Techno-economic evaluation of a floating PV system for a ...
-
Upload
khangminh22 -
Category
Documents
-
view
1 -
download
0
Transcript of Techno-economic evaluation of a floating PV system for a ...
Master of Science Thesis
KTH School of Industrial Engineering and Management
Energy Technology TRITA-ITM-EX 2019:124
Division of Heat and Power
SE-100 44 STOCKHOLM
Techno-economic evaluation of a
floating PV system for a wastewater
treatment facility
Bruno Paixão Martins
-2-
Master of Science Thesis TRITA-ITM-EX
2019:124
Techno-economic evaluation of a floating PV
system for a wastewater treatment facility
Bruno Paixão Martins
Approved
Examiner
Björn Laumert
Supervisor
Rafael Eduardo Guédez
Mata
Commissioner
Contact person
Abstract
An increasing capacity of floating photovoltaic technology (FPV) has been deployed in the past five years,
showing an alternative to countries with small surface areas to take advantage of man-made water bodies
to install solar power. Furthermore, regions where land area is primarily geared towards other uses such as
agriculture or are located in large urban centers are also a target for these installations, as land costs tend
to be a relevant share of the investment on a solar energy project. It is advertised that FPV is a more
efficient solar technology, bringing along extra benefits such as the reduction of water evaporation and
decrease in algae growth. Based on previous studies on the field as well as industry information, this thesis
aims to analyze if a floating photovoltaic project is more profitable and cost-competitive than the ordinary
ground-mounted PV for a waste water treatment facility located in Windsor, California. For that, three
scenarios are studied: Business-as-Usual, in which the extra benefits of FPV are neglected, a scenario in
which the extra benefits are included and a scenario in which external land is required to build the ground-
mounted PV. The initial approach consisted of modelling in MATLAB a 3-MW-power plant, both
floating and ground-mounted, to analyze the difference in lifetime electricity generation, capacity factors
and cell temperature. Then, an evaporation and algae growth models were proposed with the objective to
quantify the potential savings a floating PV would provide. Finally, an economic analysis was performed
to showcase the Levelized Cost of Electricity (LCOE) and Net Present Value (NPV) and used to compare
both power plants. Although, for the location chosen, on a yearly basis the cell temperature of the floating
system is on average 14.97% lower and the capacity factor is 2.54% higher, the LCOE is higher and the
NPV is lower than the ground-mounted counterpart for the BAU and extra benefits scenario. In the land-
purchase scenario, when considering extra benefits, the FPV becomes more profitable than the ground-
mounted design when over 70% of external land area is required to build the project, whereas neglecting
extra benefits, FPV becomes more profitable when 85% of external land area is needed.
-3-
Sammanfattning
De senaste fem åren har det skett en ökad kapacitet av flytande solcellsteknologier (FPV), vilket ger
alternativ för länder med begränsad yta att utnyttja konstgjorda vattendrag för att installera solenergi.
Dessutom är regioner där landområdet huvudsakligen brukas för andra ändamål, så som jordbruk eller
belägna i centrala urbana miljöer, ett mål för dessa anläggningar, eftersom markkostnaden ofta utgör en
stor del av investeringen på ett solcellsprojekt. FPV utges vara en mer effektiv solteknologi då det medför
flertalet fördelar, så som minskad vattenavdunstning samt minskad tillväxt av alger. Baserat på tidigare
studier inom området samt information från industrin, är syftet med denna rapport att analysera om ett
projekt med flytande solceller är mer lönsamt och kostnadskonkurrenskraftigt än ett vanligt markbaserat
solcellsprojekt för en avloppsreningsanläggning i Windsor, Kalifornien. För att nå syftet studeras tre
scenarier; Business-as-Usual (BAU), i vilket de tillförda fördelarna med FPV försummas, ett scenario där
de extra fördelarna ingår samt ett scenario där extern mark krävs för att bygga den vanliga markbaserade
solcellanläggningen. Det initiala tillvägagångssättet bestod av att modellera en kraftanläggning på 3 MW i
MATLAB, både flytande och markbaserad, för att analysera skillnaden i livslängd, kapacitetsfaktor samt
celltemperatur. Därefter föreslogs en modell för vattenavdunstning och algtillväxt med målet att
kvantifiera de potentiella besparingarna som en FPV skulle medföra. Slutligen genomfördes en ekonomisk
analys för att bedöma nivån på elpriset (LCOE) samt projektets avkastningsvärde (NPV), vilket användes
för att jämföra kraftverken. På årsbasis, för den valda platsen är celltemperaturen för det flytande systemet
i genomsnitt 14.97% lägre och kapacitetsfaktorn 2.54% högre, LCOE är högre men NPV lägre än det
markbaserade motsvarigheten för BAU samt scenariot som inkluderade de extra fördelarna. I scenariot
där extra mark krävs, blir FPV mer lönsamt än det för den markbaserade när det krävs mer än 70% av
externa markområden för att realisera projektet. Om de externa fördelarna försummas, blir däremot FPV
mer lönsamt först när det krävs över 85% extern mark.
-4-
Acknowledgement
I would first like to thank my family, who provided me with the best education I could ask for and gave
me continuous encouragement throughout all my education years and decisions.
Second, I’d like to acknowledge all the people that helped this thesis be as rich in information as possible.
I had the opportunity to talk with the pioneers on the development of the floating technology in the US
and for that I’m very grateful.
I’d like to thank Mr. Dale Roberts from Sonoma County Water Agency for providing me with useful
information on the operation of a wastewater facility as well as provide me with data for some of my
work.
I must express my gratitude to the people representing companies who were open to share information
with me: Ms. Coppinger and Mr. Janakes from Lake County Special Districts, Ms. Chiavari and Mr. Wolan
from New Jersey American Water, Mr. Long from Kunde Family Winery and Ms. Knego from Gundlach
Bundschu Winery.
Finally, I’d like to thank my thesis supervisor Mr. Rafael Mata for the feedback and support.
Bruno Paixão Martins
Lisbon, May 2019
-5-
Table of Contents
Abstract ........................................................................................................................................................................... 2
1 Introduction ........................................................................................................................................................13
1.1 Contextualization ......................................................................................................................................13
1.2 Objectives ...................................................................................................................................................14
1.3 Methodology ..............................................................................................................................................15
2 Literature review .................................................................................................................................................15
2.1 Solar Energy ...............................................................................................................................................15
2.1.1 Solar PV fundamentals ....................................................................................................................16
2.1.2 Types of PV layouts.........................................................................................................................19
2.2 Floating PV ................................................................................................................................................25
2.2.1 Potential Markets and Policies .......................................................................................................28
2.2.2 Panels .................................................................................................................................................30
2.2.3 Float structures .................................................................................................................................32
2.2.4 Mooring and Anchoring .................................................................................................................32
2.2.5 Cables and auxiliaries .......................................................................................................................34
2.2.6 Tracking and concentrating ............................................................................................................34
2.2.7 Environmental impacts ...................................................................................................................36
2.3 Wastewater treatment facilities ...............................................................................................................36
2.3.1 Waste stabilization ponds method ................................................................................................37
3 Case Study: Sonoma County, California .........................................................................................................38
3.1 The US scenario ........................................................................................................................................38
3.2 The role of policies ...................................................................................................................................39
3.3 The evaporation problem ........................................................................................................................40
3.4 Algal blooms ..............................................................................................................................................41
3.5 Electricity prices ........................................................................................................................................42
3.6 Defining the location and resources ......................................................................................................43
4 Modelling .............................................................................................................................................................44
4.1 Design of the power plants .....................................................................................................................45
4.1.1 Common Design Parameters .........................................................................................................45
4.1.2 Floating PV design ...........................................................................................................................48
4.1.3 Single diode model ...........................................................................................................................49
5 Economic Analysis .............................................................................................................................................52
5.1 Levelized Cost of Electricity ...................................................................................................................52
5.2 Net Present Value .....................................................................................................................................57
5.3 Evaporation................................................................................................................................................57
5.4 Algae growth ..............................................................................................................................................60
-6-
6 Results ..................................................................................................................................................................62
6.1 Design .........................................................................................................................................................62
6.2 Economics ..................................................................................................................................................65
6.2.1 Business-as-Usual scenario .............................................................................................................66
6.2.2 Extra benefits scenario ....................................................................................................................68
6.2.3 GMPV with land acquisition scenario ..........................................................................................71
6.2.4 Summary of results ..........................................................................................................................74
7 Conclusions .........................................................................................................................................................75
8 Limitations and Suggestions for future work ................................................................................................77
References .....................................................................................................................................................................78
Appendix A: CanadianSolar CS3U-350P .................................................................................................................83
Appendix B: Delta M80U inverter ...........................................................................................................................84
Appendix C: Sun path in Windsor, CA ...................................................................................................................85
Appendix D: Floating PV projects ...........................................................................................................................86
-7-
List of Figures
Figure 1: Estimated recoverable reserves for fossil fuels and yearly potential for renewables in TWy/y ....16
Figure 2: IV curves for variation on irradiance and temperature ........................................................................18
Figure 3: Effect of shading on a solar cell [6] .........................................................................................................18
Figure 4: Roof-mounted PV [8] ................................................................................................................................19
Figure 5: Premium incremental US$ per Watt of rooftop PV installed [9]........................................................20
Figure 6: PV in large corporate rooftop and carports [8] .....................................................................................20
Figure 7: Ground mounted installation on a concrete poured footing [7] .........................................................21
Figure 8: Impacts of tracking systems on land-use intensity [10] ........................................................................22
Figure 9: (a) Type of roof integrated PV, Tegola Solare roof tiles in Italy (b) Amorphous PV façade in
Albasolar Head Office, Italy [11] ..............................................................................................................................22
Figure 10: 10 MW canal top PV in Vadodara, Gujarat region, India [12] ..........................................................23
Figure 11: (a) Offshore on reclaimed land in the Kagoshima Bay, Japan [7] (b) 96kWp floating offshore
solar on shallow waters in Baa Atoll, Maldives [13] ...............................................................................................24
Figure 12: PVNB alongside road at Marano di Isera, Italy [15] ...........................................................................25
Figure 13: Common setup of a FPV power plant [16] ..........................................................................................26
Figure 14: Far Niente Winery irrigation pond FPV in California (retrieved from Google Maps) .................27
Figure 15: Percentage of reservoir area required for FPV to match hydropower dam’s capacity [16] ..........27
Figure 16: (a) Kelseyville County Waterworks 252 kWp on a facultative pond, California (b) Orlando
Utilities Commission 32 kWp, Florida [19] (c) 75 kWp on water treatment facility pond, Walden, Colorado
[20] .................................................................................................................................................................................28
Figure 17: FPV potential assessed by available man-made reservoirs ................................................................29
Figure 18: Annual installed FPV capacity [16] ........................................................................................................29
Figure 19: (a) Typical PV module assemble [26] (b) Polymeric frame PV [27] .................................................31
Figure 20: (a) Pure floats, adapted from [33] (b) Metal racks built over pontoons [32] ..................................32
Figure 21: (top) Rigid cabling mooring system anchored at the bank (bottom left) Rigid bracing (bottom
right) and spring cables [17] .......................................................................................................................................33
Figure 22: (left) Traditional dead-weight anchoring to the bottom [34] (right) auxiliary weight mooring
coupled with pulleys [35] ............................................................................................................................................34
Figure 23: Solaris Synergy solution [36] ...................................................................................................................34
Figure 24: (a) Confined tracking (b) Bow thrusters tracking [32] ........................................................................35
Figure 25: (a) Flat reflectors at the rear of neighboring panel (b) V-shaped concentration in Colignola, Italy
[32] .................................................................................................................................................................................35
Figure 26: (a) Ellis Creek Water Recycling Facility in Petaluma, California (retrieved from Google Maps)
(b) Synthetic liner being laid out [22] .......................................................................................................................38
Figure 27: (a) State Incentives for Renewables & Efficiency adapted from DSIRE Database [39] (b)
Renewable electricity installed capacity per capita adapted from [38] ................................................................39
Figure 28: 2017 PV installations by region in the USA [40] .................................................................................40
Figure 29: (a) Duckweed in a pond [42] (b) Algal bloom seen from satellite in Lake Eire near Detroit [32]
........................................................................................................................................................................................41
Figure 30: Average price of electricity to industrial consumers in the USA [45] ..............................................42
Figure 31: Wastewater treatment site coordinates: 38°31'05.0"N 122°48'10.6"W (retrieved from Google
Maps) .............................................................................................................................................................................43
Figure 32: Block diagram of the modelling .............................................................................................................44
Figure 33: Inverter power clipping [50] ...................................................................................................................46
Figure 34: Scheme representing the interrow spacing calculation .......................................................................47
Figure 35: Correlation between air and water temperature ..................................................................................49
Figure 36: Single diode equivalent circuit of a PV cell [54] ..................................................................................49
Figure 37: Specific cost of Floating PV projects below 3 MW (Compilation based on media releases and
industry information on Appendix D) .....................................................................................................................53
-8-
Figure 38: Specific cost of Floating PV projects between 3 MW and 10 MW (Compilation based on media
releases and industry information on Appendix D)...............................................................................................54
Figure 39: Specific cost of Floating PV projects over 10 MW (Compilation based on media releases and
industry information on Appendix D) .....................................................................................................................54
Figure 40: Specific cost of Floating PV projects in the USA (Compilation based on media releases and
industry information on Appendix D) .....................................................................................................................55
Figure 41: Historical water rates profile per aqueduct [60] ...................................................................................60
Figure 42: Hourly average cell temperature for both designs ..............................................................................63
Figure 43: Lifetime electricity generation and capacity factors per year .............................................................63
Figure 44: (a) Floating PV layout (b) Ground-mounted PV layout (c) High shading on the southeastern
part on winter solstice .................................................................................................................................................65
Figure 45: Sensitivity analysis of NPV varying the structural costs of the FPV ...............................................67
Figure 46: Daily evaporation rates of the chosen pond ........................................................................................68
Figure 47: Estimated mean annual evapotranspiration for the US, Sonoma County highlighted by a red
circle. Adapted from [69] ...........................................................................................................................................68
Figure 48: Algal growth rates with no cover and partial FPV cover ...................................................................69
Figure 49: Growth rate of N. oceanica with different temperature and light intensities [62] .........................70
Figure 50: Oscillatoria agardhii growth rate in fuction of temperature ..............................................................70
Figure 51: (a) Average land price per acre in California by county in 2017 [71] (b) Benchmarked lands [72]
........................................................................................................................................................................................72
Figure 52: LCOE evolution with the increase in external land required compared to FPV’s LCOE ...........73
Figure 53: NPV evolution with the increase in external land required compared to FPV’s NPV .................73
Figure 54: LCOE results for the analyzed scenarios .............................................................................................74
Figure 55: NPV results for the analyzed scenarios ................................................................................................75
-9-
List of Tables
Table 1: Energy yield (kWh) and specific yield (kWh/kW) for different tilt angles .........................................46
Table 2: Cost breakdown of ground-mounted and floating PV ..........................................................................56
Table 3: Function parameters for the species evaluated .......................................................................................61
Table 4: Common design parameters for FPV and GMPV .................................................................................62
Table 5: Constant economic parameters .................................................................................................................66
Table 6: BAU scenario results ...................................................................................................................................66
Table 7: Extra benefit scenario results .....................................................................................................................71
Table 8: Floating PV projects around the world ....................................................................................................86
Table 9: Sources for the FPV projects costs ...........................................................................................................88
Nomenclature
Alternate Current AC
Airport/Larkfield/Wikiup Sanitation Zone ALWSZ
Business-as-Usual BAU
Biochemical Oxygen Demand BOD
Balance of the System BoS
Building Integrated Photovoltaic BIPV
Capacity Factor CF
Concentrated Solar Power CSP
Direct Current DC
Energias de Portugal EDP
Engineering, Procurament and Construction EPC
Ethylene-vinyl acetate EVA
Feed-in-Tariff FIT
Floating Photovoltaic FPV
Ground Coverage Ratio GCR
Ground-mounted Photovoltaics GMPV
High Density Polyethylene HDPE
The International Electrotechnical Commission IEC
International Organization for Standardization ISO
Investment Tax Credit ITC
Levelized Cost of Electricity LCOE
Maximum Power Point Tracker MPPT
Nominal Operating Cell Temperature NOCT
Net Present Value NPV
National Renewable Energy Laboratory NREL
The Organisation for Economic Co-operation and Development OECD
Pacific Gas & Electric Company PG&E
Power Purchase Agreement PPA
Production Tax Credit PTC
Photovoltaic PV
Photovoltaic Noise Barrier PVNB
Renewable Market Adjusting Tariff ReMAT
Renewable Portfolio Standard RPS
Rijksdienst voor Ondernemend Nederland RVO
System Advisor Model SAM
-10-
Sonoma County Water Agency SCWA
Solar Energy Research Institute of Singapore SERIS
Standard Test Conditions STC
United Kingdom UK
United States of America USA/US
Water Regulations Advisory Scheme WRAS
Waste Stabilization Ponds WSP
PV sizing
Diag Diagonal shading distance
L Length of the panel
β Tilt angle
αs Solar altitude angle
γs Surface azimuth angle
d shading distance
Tamb Ambient temperature
Vwind Wind speed
Twater Temperature of the water body
μ Estimated minimum stream temperature
α Estimated maximum stream temperature
βair Air temperature at inflection point
γ Steepest slope of the function
I Current
Il Light current
Io Reverse saturation current
Rs Series resistance
Rsh Shunt resistance
n Diode ideality factor
V Voltage
a Ideality factor
Ns Number of cells in series
Vt Thermal voltage
G Irradiance
αi Short circuit current coefficient
Tcell Cell Temperature
k Boltzmann constant
Ebg Silicon band gap energy
Voc Open circuit voltage
Isc Short circuit current
Imp Maximum power point current
Vmp Maximum power point voltage
Pdc Maximum DC power
DClosses DC losses
PACstring AC string power
PDCstring DC string power
AClosses AC losses
ηinverter Inverter efficiency
Npanels Number of panels
DCpower Power capacity
-11-
Pmp Maximum power
CF Capacity factor
Economical
LCOE Levelized Cost of Electricity
It Initial investment costs
OMt Operation and Maintenance costs
Ft Fuel costs
Et Energy generated
r Discount rate
t time
Evaporation model
ETo Evapotranspiration rate
Tair Mean air temperature
u2 Wind speed at 2 meters height
Rn Net radiation flux
Gsoil Soil heat flux
Δ Slope of the vapor pressure-temperature curve
γp Psychrometric constant
es Saturated vapor pressure
ea Actual vapor pressure
αalb Albedo
σ Stefan-Boltzmann constant
Rsi Measured solar radiation
Rso Clear sky solar radiation
Tmax Maximum daily temperature
Tmin Minimum daily temperature
Tdew Dew temperature
Rsa Extra-terrestrial solar radiation
dr Relative distance between Earth and sun
ωs Sunset time angle
φ Latitude
δ Solar declination
J Day of the year
Gsc Solar irradiance constant
λvap Heat of vaporization
P Pressure
Cpair Specific heat capacity of air
uh Wind speed at certain height
h Height
Algae growth model
λ Wavelenght
QF Quantum flux of light
μg Algae growth rate
μm Maximum specific growth rate
-12-
μopt Optimal growth rate
i Rate between light intensity and optimum light intensity
Iopt Optimal light energy
Φ Model parameter
w Algae coefficients
u Algae coefficients
γm Algae coefficients
αg Algae coefficients
T Temperature
To Temperature where algae coefficients are zero
Topt Temperature where algae coefficients are maximum
Tmax Maximum temperature before growth inhibition
Tmin Minimum temperature before growth inhibition
-13-
1 Introduction
1.1 Contextualization
Social, political, economic and environmental conjectures have frequently dictated the pace of which
humankind pursued new energy alternatives. Take the oil crisis of 1973 for example, caused by an
embargo of the Organization of Arab Petroleum Exporting Countries, which caused an international oil
price shock, when many of the countries had the energy matrix highly oil-dependent. This led to many
countries being affected economically, driving the research of new energy technologies. Brazil, greatly
affected in the transport sector, started its program of ethanol-run engines which led to a successful
introduction of a new and cleaner combustible fuel to the market. Denmark which had more than 80% of
its energy needs covered by imported oil, started pushing policies and mandates that set the grounds for
what the country has become, a leader in wind power per capita. The US pushed for new policy packages
that deregulated prices and boosted federal research funds which led to breakthroughs of new
technologies such as solar and shale, but also on coal power plants. Although not designed to reduce
carbon emissions, these alternative measures led to a reduction in the growth rate of carbon emissions
from 4.5% before 1973 to 0.4% after 1973. Globally speaking, had carbon emissions kept with the pre-
1973 pace, the 2012 emissions would have been three times the actual level [1].
Forward in time, the always evolving energy landscape has established itself another important event: the
Paris Agreement which was ratified by 184 parties. The agreement proposes a long-term objective to keep
the increase in global average temperature below 2 °C pre-industrial levels and efforts to limit it until 1.5
°C, which should ease the effects of climate change. This initiative is driven by environmental concerns
and clearly impacts all other areas of society, as countries will have to adapt socially, but mainly,
economically to changes in order to reach the proposed goal.
Although energy alternatives were always in development, past decisions created a path dependency that
locked in fossil fuels as preferred option to invest. Therefore, all the economic sectors have relied on fossil
fuel for decades: industry, transport, residential and electricity have used oil, gas and coal in a way or
another as energy sources. To add insult to injury, energy demand is projected to increase 28% by 2040,
with more than half of this increase due to China and India [2]. However, with the increased
environmental concern following Kyoto protocol, its amendments and the Paris Agreement, efforts have
been shifted to sustainable alternatives to fulfill this demand. Many technologies have evolved such as
EV’s, hydrogen power, biofuels, wind, solar and nuclear and many others are in development such as tidal,
wave and storage systems. The famous quote by the former ministry of oil of Saudi Arabia, Sheikh
Ahmed Zaki Yamani, in which he stated that “The Stone Age came to an end, not because we had a
lack of stones, and the oil age will come to an end not because we have a lack of oil.” [3] may not look like
a utopia after all.
An evident development has occurred in the electricity generation business, as renewable energy
technologies have flourished, specially wind and solar power. With its demand also expected to increase
on average 1.9% and 1.0% per year on non-OECD and OECD countries, respectively, wind and solar will
become major contributors to installed capacity [2]. Countries like Denmark, Portugal and Costa Rica, to
name a few, have accomplished several days being powered only by renewable energy sources, being the
latter able to run only on renewables for 300 days in 2017 [4].
Behind wind power, the spearheading renewable energy technology, solar power has demonstrated
throughout the years its potential to be the energy of the future. Its flexibility to be installed in many
locations (rooftop, ground-mounted, building integrated and floating over water) and declining costs has
attracted the attention of many sectors as a possible solution to offset the consumption from the grid to
in-house and take part in the upcoming energy transition. Although the simplicity of the technology, many
factors influence in the decision to install such systems. Large scale utilities and residential consumers have
-14-
already started to implement solar power as an alternative, following the need to abide to certain policies
or as an economically benefit solution.
Meanwhile, the industry sector sits in a stalemate, as many rely on the grid to buy electricity to power their
manufacturing processes, clearly set by another path dependency created by the way energy is
commercialized. If the country produces electricity utilizing fossil fuels, industries will undeniably be part
of this matrix. Under the scope of the recent changes and agreements, in order to take part in the energy
transition, solar generation appears as a leading option. However, many industries have a limitation of
space or not enough area on building roofs to invest in a solar photovoltaic. Yet, these same industries
may also have ponds (effluent treatment, irrigation, storage, etc.) which are areas that could be used to
harness electricity, making floating photovoltaics an interesting alternative. Wastewater treatment facilities
make up for a great potential customer for this technology as good part of the area is composed by ponds.
The United States, as an economic powerhouse, a leading country in installed solar capacity and, to certain
extent, policy friendly, is a great location for the development of such projects, since it could benefit from
both the market and technological development that floating photovoltaics have shown in other parts of
the world, especially in Asia.
Given the multiple challenges stated above, several solar technologies offer potential to power the
industry. In order to assess if a floating solar power plant will be adequate for a wastewater treatment
facility, it is of interest to understand the different technical aspects and quantify all the possible knock-on
benefits that may improve the attractiveness of such alternative over a ground-mounted (GMPV) or roof-
mounted one, for example. This topic is particularly interesting considering the ongoing change on
policies landscape that may help promote such technologies.
1.2 Objectives
The thesis has the main objective to compare a floating photovoltaic system to other mounting layouts
and point out which project is more attractive in terms of economic benefits for a wastewater treatment
facility located in the USA to cover its power demand and eventually sell the surplus of energy to the grid.
The technology will be analyzed through the scope of:
a. Technical analysis: compare technical aspects between a grid connected floating PV solution to a
ground-mounted PV;
b. Quantification of qualitative benefits of a floating PV that otherwise wouldn’t exist if the system
were to be mounted another way;
c. Economic analysis: compare LCOE and NPV from floating PV to the other selected alternative,
considering three different scenarios.
a. Business-as-Usual (BAU) scenario: FPV without quantified extra benefits and ground-
mounted PV;
b. Extra benefit scenario: FPV with quantified extra benefits and ground-mounted PV;
c. External land purchase: FPV versus ground-mounted PV requiring extra land to build
the system.
Following this rational, it is expected to be able to compare both technologies through same-basis
quantitative indexes which should give a suggestion to which alternative is more attractive to be
implemented in the chosen location.
Ultimately, this work is proposed not only as a graduation assessment instrument, but also as a study that
could be a starting point for development of future relevant work.
-15-
1.3 Methodology
After a brief introduction to contextualize the problem, the research work will be structured as follows.
Section 2 will be reserved to describe most commercially available mounting design that integrate
photovoltaics. It will be given extra attention to details when reporting about floating photovoltaic (FPV),
as it is the main technology to be studied here. At the end of the same section, a brief review will be given
on the type of business this thesis is focused on which are wastewater treatment facilities. Section 3 will
get into the details of the selected location of the research, namely the resources availability and all the
factors that led into choosing it. With all the background set, Section 4 will be dedicated to the technical
modelling of the FPV. Section 5 will be dedicated to the economic analysis of the system of the previous
chapter, alongside the quantification of the common suggested qualitative advantages of FPV. Section 6
will be devoted to the comparison of the power plants proposed to the wastewater facility taking into
account both technical and economic analysis. Finally, the conclusions chapter will be presented
highlighting the most important results found, limitations and suggestions for future work developments.
2 Literature review
The following chapter contains multiple information that are of interest to the development of this thesis
work. The theoretical framework will cover technical aspects of PV fundamentals and equipment, since
this is the ground zero of the design of such installations and will be important for the technical
simulation and economic evaluation. The market and the policies will also be analyzed, with special focus
on the US.
Finally, a brief review of wastewater treatment facilities will be done in order to understand the processes
and physical setups, thus, grasping the location that could be suitable to install the proposed PV solutions.
2.1 Solar Energy
The sun is one of the most important natural resources and is responsible for many biological and
energetic processes on the planet. Figure 1 shows the potential power that can be harnessed by each of the
renewables sources in the world and also the estimated power capacity recoverable by proven reserves of
fossil fuels, on a 2015-year basis [5].
-16-
Figure 1: Estimated recoverable reserves for fossil fuels and yearly potential for renewables in TWy/y
There are many ways to convert solar radiation to useful energy for human needs. In the field of power
generation two main categories of technologies exist: thermal and photoelectric effect conversions.
Thermal conversion is related to Concentrated Solar Power (CSP) technologies. Concentrating mirrors
converge the sun rays into a fluid that is superheated and utilized as heat source for a thermodynamic
cycle (usually Water Rankine Cycle, although Organic Rankine Cycles may also be used). Two main
technologies are widely used in CSP facilities: parabolic trough concentrator and solar tower. Photoelectric
effect is related to solar photovoltaic cells and is a technology that widely spread. The main concepts will
be discussed in the sub-section below.
2.1.1 Solar PV fundamentals
2.1.1.1 Components
Solar PV is one of the considered renewable energy sources that have the most potential for electricity
generation, as it has been previously observed in Figure 1. Currently, the way solar energy is converted to
electricity is through the use of photovoltaic panels that are made of a series connection of solar cells
made of semiconductor materials such as, but no limited to, silicon. These cells generate electricity by
exploiting the photoelectric effect, which consists in absorbing the energy of a photon, particle that
represents a quantum of light or electromagnetic radiation, and raising an electron to an excited
energy state. When connected to an external load, the circuit is closed and the electron is captured
generating direct current (DC) current. Since most of the household electrical equipment are powered
by alternate current (AC), an inverter has to be used to convert it into useful current. The usual
nomenclature and equipment used in a photovoltaic installation are summarized below.
• PV cell: made of semiconductor material, able to generate electricity when light is cast and an
external load is connected. It represents the smallest unit of the system;
• PV module: is the interconnection of a series of PV cells. Common crystalline silicon PV modules
have between 60 to 72 cells in series, however depending on the use, the number of cells wired in
series might differ and the size of the panel as well. Different types of technologies exist such as
-17-
poly- and monocrystalline silicon, amorphous silicon, thin film, dye-sensitized, organic film and
others;
• PV array: is the wiring together of multiple PV modules that may be connected in series, forming
a string of panels and also in parallel. PV arrays may range from single panels on rooftops, for
example, to utility scale MW projects having several strings of panels;
• Inverter: electronic device that converts the DC current generated by a PV array to AC current to
be utilized by electrical equipment or fed into the grid. Inverters can be connected to multiple
arrays and are called central, or may be connected to strings, called string inverters or connected
to every panel, which in this case are called microinverters. The last gives a better control of the
output of each individual panel, but has increased costs as a drawback;
• Batteries: as solar power is intermittent and power consumption may not match the time of
power generation, batteries can be used to store the energy generated and discharged when
consumption starts;
• Balance of the System (BoS): All other components that make up the PV installation such as
cables, wires, switches, automatic protection equipment, fuses, racking, mounting structure and
others.
2.1.1.2 Design and operational aspects
Solar panels are characterized by a number of parameters, measured in a set of controlled variables called
Standard Test Conditions (STC). This scenario represents the solar cell under direct normal irradiance of
1000 W/m², ambient temperature of 25 °C and 1.5 Air Mass atmosphere. The main parameters are briefly
described below:
• Nominal Maximum Power (Pmax): nameplate capacity of the module. Under STC, the power
output should be equal to Pmax;
• Maximum Power (Pmp): the pair of current and voltage that delivers the maximum power under
certain irradiance and temperature conditions;
• Maximum power point voltage (Vmp): the voltage of the module under the maximum power
point;
• Maximum power point current (Imp): the current generated under the maximum power point;
• Opern circuit voltage (Voc): the voltage when no current exists;
• Short circuit current (Isc): the current when no voltage is applied on the terminals;
• Module Efficiency: represents the ratio of the power output at the maximum power point under
certain conditions with respect to the STC.
On the real environment, irradiance and temperature vary throughout the whole day and impact the
output of the solar panel. With decreased irradiance, output current is greatly reduced and voltage drops
slightly. Ambient temperature affects the cell temperature, which has been widely investigated by several
authors and usually described by the Nominal Operating Cell Temperature (NOCT) model that
accompanies the manufacturer datasheet. With an increase in temperature, voltage may drop greatly,
whereas current may increase slightly. The behaviors described before can be seen in Figure 2 retrieved
from datasheet of the panel on Appendix A.
-18-
Figure 2: IV curves for variation on irradiance and temperature
For every instant, irradiance and temperature are varying which means that the PV module will have a
unique IV curve and a new maximum power point. If this point is not tracked, the PV panel will work on
a fixed condition which isn’t the optimal and, therefore, the output power will not be maximized. For this
problem, inverters are normally equipped with maximum power point trackers (MPPT) which are
algorithms embedded into the equipment or charge controllers that compare the instant voltage to the
grid or battery voltage and converts it to the maximum voltage.
Ideally the operation of a solar PV system should occur without shading, which directly affects the output
current of the module. When a section of the panel is shaded, certain cells will be fully illuminated while
others will receive a different irradiance and this will represent another unique IV curve (Figure 3),
decreasing the output proportionally to the amount of shading. A poor illuminated cell acts as a blocker to
the flow of current from the fully illuminated ones, which leads to a reverse bias operation across the
shaded cell and power dissipation. This can create hotspots which can cause cell and glass cracking,
melting and degradation. For this issue, single or sets of solar cells are often coupled in parallel with
bypass diodes, which conduct the current when a shaded cell is reverse biased.
Figure 3: Effect of shading on a solar cell [6]
Another important aspect of the design phase is the choice of inverter. The interconnection of PV
modules in series creates a string voltage, which is given by the sum of all the panels voltages. When these
strings are connected in parallel the current will be increased, as it will be the sum of the currents of every
-19-
string. These parameters are important when sizing the inverter, which will have to withstand the load
provided by the arrays. For large utility scale solar plants, it is common to see central inverters whereas for
commercial and residential installations, string inverters or microinverters can be used.
2.1.2 Types of PV layouts
Several types of PV mounting have been developed to adequate to many applications. In the following
sections the main commercial configuration will be presented, focusing on the advantages and
disadvantages of such systems. Furthermore, if possible, the applications on the industry sector will be
shown as this thesis focus on a solution for wastewater treatment facility that is part of this sector.
In addition, benefits or secondary applications that normally are overlooked will be highlighted. This is
relevant, since this work will analyze the extra benefits of FPV and thus the analysis can be discussed over
these topics as well. The floating PV section will be separated into sub-sections as to analyze more
thoroughly the State-of-the-Art of the components, since this layout hasn’t been around as much as other
applications.
2.1.2.1 Roof-mounted
One of the most popular setup that can be found is the roof-mounted PV. The installation consists of PV
panels covering partially -or totally- the roof of a building, which is a great advantage in big cities where
land is restricted. Given the average size of the homes and roofs, those installations are usually comprised
between 5-20 kW for residential buildings and peak up to 100 kW for commercial buildings [7].
In order to keep costs at the minimum and optimize spacing, the panels are usually installed following the
inclination of the roof. Therefore, rooftop PV are bound to the inclination and direction of the roof,
meaning that, given no structural modifications, the installation may not be in the optimal layout to
maximize energy yield, i.e., northern hemisphere westward facing home roofs with an inclination below
the latitude angles. Another challenge is that in urban settings roofs are close to several obstacles such as
chimneys, satellite dishes and trees, which may create shading in certain times of the day. A typical roof-
top photovoltaic can be seen in Figure 4.
Figure 4: Roof-mounted PV [8]
It is argued that rooftop PV, besides the obvious benefit of microgeneration and an extra layer of
protection on the roof against weather tear, may also increase the real state property value. In fact, it was
found that California home buyers are willing to pay a premium of up to US$4/W or US$15,000 for a
-20-
house with a system of 3.6 kW average capacity [9], which can be seen in Figure 5, and even greater values
are paid by non-California homes.
Figure 5: Premium incremental US$ per Watt of rooftop PV installed [9]
Large corporations have also started to install solar PV on the rooftop of their buildings. This is normally
seen in the headquarter buildings or in large horizontal roof warehouses or distribution centers. Solar
installations may also be used as shading structures for carports in parking lots. Both layouts are observed
in Figure 6.
Figure 6: PV in large corporate rooftop and carports [8]
As of 2017, several North American corporations have installed large capacities of solar PV, with the retail
giant Target leading the way with 422 installations accumulating 204 MW of installed capacity. Main
drivers may lie in demonstrating environmental responsibility by committing to sustainability which
strengthens brand. Target, for example, has reported in 2017 an objective to power 100% of all its
-21-
operations with renewable energy. Another reason may lie in the financial stability, since these buildings
consume a large amount of electricity and any price variation may cause large budget distortions. In the
case of the US, financing was made easy with the spread use of Power Purchase Agreements (PPA) and
the federal Solar Investment Tax Credit (ITC) that companies use to lower the financing costs [8].
2.1.2.2 Ground-mounted
Another common configuration is the ground-mounted solar power plants. This configuration is normally
chosen for the deployment of large-scale utilities and it is one of the best choices if area is not a
restriction. The mounting systems can be fixed on many different types of structures such as single poles
inserted directly into the ground, inserted through a concrete pouring foundation (Figure 7) or mounted
on concrete ballasts when digging is not possible. These structures should be resistant to wind loads as to
prevent uplifting of the panels. Rack mounting, using steel pipe racks for example, is commonly used for
placing panels close to the ground or in horizontal rooftops.
Figure 7: Ground mounted installation on a concrete poured footing [7]
One of the main advantages of ground mounted PV is the ability to install solar tracking systems.
Tracking the sun means that the energy output will be greater and the energy yield per square meter will
also increase. Different types of tracking exist as follows:
• Single axis: rotates in a single direction and can be horizontal (east-west or north-south), vertical,
tilted and polar (parallel to the earth axis);
• Dual axis: always point directly to the sun and can be tip-tilt and azimuth-altitude oriented.
Although it is possible to increase energy yield, it comes with an increase in capital costs, maintenance and
more area per unit of capacity [10]. One has to analyze which is the benchmark parameter to be
considered in order to install a solar facility with or without tracking. The following figure exemplifies a
study on the increase or decrease of land-use intensity (acres/GWh/year) for seven different locations in
the US.
-22-
Figure 8: Impacts of tracking systems on land-use intensity [10]
Not only large utilities companies are producing clean energy with large scale photovoltaic facilities. Big
corporations have also adhered to this market by installing on rooftop PV, as it was in section 2.1.2.1, but
also in large areas when available. Another options that has emerged is the off-site corporate procurement
that drove tech-companies like Facebook, Amazon and Microsoft to contract projects outside their
properties to offset electricity demand. Companies may benefit from this type of scheme by two means:
either by physical PPAs which the company consumes the electricity or virtual PPAs which the company
benefits from the electricity sold to the market [8].
2.1.2.3 Building integrated photovoltaics (BIPV)
Building integrated photovoltaics are supposed to be an integral part of the final building. The panels or
tiles are not only generating electricity but also represent part of the structure. Correctly installed BIPV
may also offer other benefits such as shading, thermal insulation and noise protection [11]. The
technology is mainly focused in new buildings, but may also be used in refurbishing of old buildings,
through façade or roof integration. Different types of integration exist: roof tiles, roof shingles, semi-
transparent and opaque wafers, frameless panels and many others. A roof tile and a façade example can be
seen in Figure 9.
(a)
(b)
Figure 9: (a) Type of roof integrated PV, Tegola Solare roof tiles in Italy (b) Amorphous PV façade in Albasolar Head Office, Italy [11]
The market leader technology for BIPV is crystalline silicon, but thin film and amorphous silicon have
increased the market share and are frequently used, since buildings are often under partial shading
-23-
conditions and these technologies are less affected by shading than their crystalline counterparts. Another
advantage is the lower price per square meter (m²) compared to the rigid Si-wafer modules [11]. If
efficiency were to be the number one factor on the choice of technology, other types of solar panels could
be selected. Therefore, it can be understood that BIPV are chosen mainly due to the aesthetical
functionality. For example, it can be seen the building from a solar company in Figure 9 (b), that clearly
identifies itself with the product or service by utilizing amorphous silicon PV panels as part of the office
façade.
2.1.2.4 Canal top
Given the flexibility, the know-how of PV structures and the lack of land in some regions, the canal top
solution was created. Canals may be natural rivers or man-made hydraulic canals used to supply water for
consumption or irrigation. India has been in the forefront of such installations and pretends to harness up
to 100 MW of solar energy through canal top PVs [12]. Figure 10 exemplifies one of the few installations
that are running up to date.
Figure 10: 10 MW canal top PV in Vadodara, Gujarat region, India [12]
Canal top installations show some benefits for water operation companies, as it can be installed near
irrigation pumps supplied by such canals, which may be located in remote locations. This results in a
decrease on transmission and distribution losses.
Other benefits are related to water quality. Such structures create a shade over the water which can reduce
evaporation up to 9 million liters per MW installed a year [12], especially in dry regions such is the case of
some locations in India. Algae growth may also be slowed down as shade also prevents microorganisms to
perform photosynthesis. Uncontrollable algae growth may result in clogging irrigation and transport pipes
as well as pumps, causing loss in service and increased maintenance costs. If the water is used for drinking,
it may become undrinkable.
Although many benefits can be related, some disadvantages lie in these types of installation. Costs are a
main concern as the mounting structure may need wind breaking structures [12] as to curb the appearance
of a region of higher air flow, similar to a wind tunnel, which poses a risk to the panels. Following the
canal path would also mean that sometimes the panels will not be in the optimal layout. Also, depending
on the layout of the structure, the access to maintenance of the PV plant may be difficult and could
interfere in the removal of debris and residues that accumulate along time in the canal. Finally, locations
-24-
that transformed itself in natural habitats of aquatic animal species and plants could face public rejection
when considered for the development of such projects.
2.1.2.5 Offshore
While offshore wind power plants are trending, deep water solar offshore power still hasn’t become a
reality. The structure of such concept needs to withstand harsh environments, such as corrosive salty
water, as well as endurance with waves. The mooring structure needs to be robust, otherwise the floating
panels may find themselves broken and sailing in open ocean. Floating structures already exist but are
designed to lakes and inland waters that do not have high waves.
(a)
(b)
Figure 11: (a) Offshore on reclaimed land in the Kagoshima Bay, Japan [7] (b) 96kWp floating offshore solar on shallow waters in Baa Atoll, Maldives [13]
To date no offshore power plant has been built on deep ocean waters. Figure 11 (a) shows a PV power
plant installed in reclaimed water in the Kagoshima Bay in Japan, but the panels sit on land projected into
the ocean and do not have to withstand the same environment as a floating structure in the open ocean.
On the right-hand side, a floating PV is located in shallow waters, constituting tropical lagoons, which
have smaller waves up to 0.5 m [13]. These types of projects might be of interest of small islands that are
in outlying regions of the globe and do not have enough ground or rooftop to be covered with
photovoltaics. These islands normally import fuel and are very dependent on mainland countries.
However, some of the near-shore environments might be sensitive to human action which can be a
limitation to the deployment of the technology.
The first project on offshore solar farm on deep waters, about 15km from the coast of The Hague, is
being designed and built by a consortium of 6 companies financed by the Dutch Enterprise Agency
(RVO). It will first start with a pilot of 30 m², but will ultimately reach 2,500 m² by 2021 [14]. The biggest
advantages of a deep-water system will lie in the large area availability and the possibility to interconnect
with offshore wind farms, not only in the transmission of electricity but also to the robust mooring of
these structures. Also, the cooling effect of the water may improve efficiency of the photovoltaic panel
and increase energy yield.
An interesting fact that relates the ocean with photovoltaics even more is that Magnesium Chloride, found
in ocean waters, that it is used to produce bath salts and tofu for example, could be the substitute of
Cadmium Chloride, a highly toxic and expensive component, used to produce Cadmium Telluride solar
cells [7].
-25-
2.1.2.6 Other layouts and applications
With the popularization of the solar technology, many lower load applications started to be powered with
photovoltaics. Calculators, watches, street lightning, offshore navigation aid lighting, highway directional
signs, other measurement equipment and even satellites can be found being powered by solar energy.
Just like BIPV, some applications benefit more than the energy generation. The so called Agrivoltaics is
the integration between a solar power plant with a crop that can adapt to the shading provided by the
panels. Cathodic protection is also another application that PV may be used, as metallic structures that
have been buried, for example, may be able to withstand more time before undergoing corrosion when
provided with DC current. Photovoltaic noise barrier (PVNB) is another example of the panels being used
with a primary function that is different than generating electricity. In the highway near the town of
Marano di Isera in Italy, which can be seen in Figure 12 below, 3944 photovoltaic modules were installed
as barriers and are capable of producing up to 690 kWh per year of electricity.
Figure 12: PVNB alongside road at Marano di Isera, Italy [15]
2.2 Floating PV
Floating photovoltaic or nicknamed “Floatvoltaics”, is normally referred to installations that are placed
over water bodies. Submersed floating PV also exists but will not be covered in this review, as the majority
of installations consist of testbeds. The water basins chosen for installing FPV are mainly artificially built,
such as irrigation ponds, lakes, abandoned & flooded mine pits and other non-utilized water reservoirs.
The main advantage is related to land use, places where space is scarce, costly and unfit for ground-
mounted systems might benefit from these installations. Artificial ponds, which normally take up large
spaces, serve a purpose but disable other applications on that same area. Joining both facts together make
floating PVs a great solution.
On a technical side, the main advantage is related to the increase in efficiency of the panels, since there is a
cooling effect that reduces cell temperature, which directly affects energy yield. Moreover, FPV is believed
to be less prone to soiling of dust and shading, as some power plants are placed in the middle of ponds
without obstacles around. Therefore, cleaning of the panels is also made easier since there is water
available at the same site. Installation and decommissioning are also simpler since the structures are only
fixed through anchoring which may be easily removed, unlike foundations of ground-mounted power
plants.
-26-
Other potential benefits lie in more complex matters. Decrease in algae growth is pointed out as a possible
environmental benefit [7], which is especially important for wastewater treatment ponds to maintain the
stored drinkable or irrigation water and also to prevent clogging of pumps and pipes. This is suggested by
the less light reaching the water body and decreasing the rate of photosynthesis. Also related to water,
decrease in the rate of evaporation is another advertised benefit, due to the shading of the panels on top
of the pond, just like observed in canal top and offshore installations. Both benefits may represent
additional savings for companies that sell water, like wastewater treatment facilities, but until now it has
been yet to be researched. Figure 13 below depicts a traditional installation of a floating PV plant.
Figure 13: Common setup of a FPV power plant [16]
As observed, the structure is comprised of PV modules connected to a combiner box and, in this case, a
central inverter which is also on floating structure. This is not always the case, as small and shallow ponds
can have an inverter placed on shore. The biggest difference here is the floating structure and the mooring
and anchoring system, which will both be discussed further in the following sub-sections.
Although the technology isn’t novel, the occurrence of FPV plants are very recent. The first installation
was built for research purposes in Aichi, Japan [16], which ended up being one of the countries that has
developed and installed much of the today’s capacity. The first commercial installation, however, was
power plant installed in Far Niente Winery in Napa Valley (Figure 14), California in April of 2008. Since
land is an important and expensive asset for wineries, floating projects are a great solution to generate
clean energy. The winery became an example of sustainability in the production of wines, has won awards
of Solar Winery and is a leading sustainable wine production company. The awards came as a brand-
strengthening image but the FPV served an important purpose. Figure 14 shows the 2,292 modules power
plant rated 478 kWp that is divided between land and water. The ground-mounted PV utilized 1.2 acres of
land while the floating one occupied 0.75 acres of flooded land, which turned out to save US$ 150,000 in
revenue of wine being lost annually if the whole installation would’ve been placed on land only [17].
-27-
Figure 14: Far Niente Winery irrigation pond FPV in California (retrieved from Google Maps)
FPV also allows integration to many other businesses. Just like ground mounted PV can be used to
provide shading for cultures (“Agrivoltaic”) that excel on less light, FPV can be integrated to aqua culture
and fish farming. However, much interest has been drawn to the integration to hydropower dams, which
floods a huge amount of land that otherwise could be used for something else. Installing floating solar can
increase the output of hydropower plants but, most importantly, allows it to operate in peak-mode, when
there’s low availability of water due to droughts. In addition, power plants that cannot maintain full
generation capacity may also benefit of the offset generation provided by solar power [18]. Moreover, the
proximity to the already existing grid connection and transmission lines avoids the investment in new
infrastructure and also increase the rate of utilization of such lines. When there’s enough availability of
energy being produced, it may be used to pump water back to the dams. Figure 15 exemplifies the amount
of area that would need to be covered with photovoltaics in order to match the same capacity as the
hydropower plant.
Figure 15: Percentage of reservoir area required for FPV to match hydropower dam’s capacity [16]
One of the main drawbacks of floating solar on dams are costs, highly influenced by the mooring and
anchoring that need to withstand large water level differences, wind and wave loads that are way superior
than irrigation ponds for example. Energias de Portugal (EDP) installed the first pilot project in the world
in the pumped storage dam of Alto do Rabagão, Portugal, that was able to withstand intricate conditions:
along with strong winds, waves up to 1-meter-high, water variations of 30 meters in a soil bed of 60
meters’ depth. This impressive feat came with a price, taking up to 20% of the total costs of the 220 kW
project [18].
Out of the many projects that exist, several of them are placed in water holding ponds that may be from
wastewater treatment facilities, farms or private ponds. These ponds are smaller than hydropower dams
-28-
and pose less of a challenge in anchoring as there is less wave loads and changes in water level.
Wastewater treatment facilities may enjoy the full package of benefits from installing FPV, such as
reducing evaporation of the water that was just treated to be commercialized, as well as preventing
contamination from algal bloom, with the clear benefit of generating its own electricity. Depending on the
size of the pond, the design and the coverage area, the power utilized by the treatment facility may be
completely covered. Some of the installations found in the USA can be observed in Figure 16.
(a)
(b)
(c)
Figure 16: (a) Kelseyville County Waterworks 252 kWp on a facultative pond, California (b) Orlando Utilities Commission 32 kWp, Florida [19] (c) 75 kWp on water treatment facility pond, Walden, Colorado [20]
As FPV becomes more popular, several wastewater treatment facilities are starting to assess the feasibility
of this technology. As costs are still developing, several projects are below the 100 kWp. Nonetheless,
many ways to finance the installation exist, as it was the case of the town of Walden, Colorado. The town
has signed a performance contract with the supplying company, Johnson Controls, which assures the
energy savings provided by the FPV and is expected to save up to US$ 10,000 per year and cover the
energy demand of the facility in certain seasons. The simple payback is estimated in 20 years [20], which is
around the lifetime of the panels, thus enabling the project.
2.2.1 Potential Markets and Policies
Floating photovoltaic has its potential linked to the availability of water bodies distributed around the
world as well as other factors linked to costs, public policies and environment concern. The Solar Energy
Research Institute of Singapore (SERIS) has calculated the potential of covering 1%, 5% and 10% of a
number of assessed freshwater man-made water bodies, which can be seen in Figure 17. On a
conservative approach this analysis resulted on 404 GWp of solar power potential. Many of the existing
projects can be seen covering almost the whole water body, but these are normally freshwater ponds that
do not have large extensions. Great part of the area available will come from large kilometer-squared
hydropower dams. Doing a simple calculation, recalling Figure 15, covering 1% of Aswan Dam in Egypt
with 1.6 m² PV panels of 230 W will mean setting up almost 32 million panels on water with a capacity of
around 7 GW, which is almost 5 times the size of the largest ground-mounted PV facility in Tengger
Desert, China.
-29-
Figure 17: FPV potential assessed by available man-made reservoirs
From the figure above, it can be observed great potential in Africa, Middle East, Asia and North America.
Asia has a huge potential in China, as the country holds the number one spot on solar installed capacity
per year, highest production of solar panels and one of the latest and largest floating PV power plant (40
MW in Huainan City), while countries like Japan, Taiwan and South Korea will continue develop the
market on a limited surface. Huge potential is also found in North America, as it holds a great share of
population and a developed water treatment and irrigation system, as well as several hydropower plant
dams. In a recent study, 24,419 man-made waterbodies were found to be suitable for floating PV
installations, whereas covering only 27% of the available surface would be capable of producing nearly
10% of the current country’s generation [21]. These numbers may account for the over 8,000 wastewater
treatment ponds that exist and could be used to place FPV [22].
In fact, in just a few years, floating PV increased exponentially, reaching up to 1 GWp of installed
capacity, having a limited number of countries with larger scale installations such as Japan and China.
Nonetheless, the high increase is an impressive feat for a technology that was first deployed in 2007 as a
test and only in 2008 with the first commercial installation. Figure 18 shows the evolution of installed
capacity in the past decade.
Figure 18: Annual installed FPV capacity [16]
As a renewable energy solution, FPV is entitled to the same benefits regular photovoltaics have. Clearly,
this depends on the will of government and private institutions to work the best approach to develop the
technology. It is observed that higher Feed-in-Tariffs (FIT) in Asian countries, specially Japan and Taiwan,
-30-
have helped an accentuated development of installations. Moreover, these island countries have area
limitation which makes FPVs a perfect solution to be placed on ponds that are consuming large parts of
the surface.
After the incident in Fukushima, the policy landscape in Japan has changed and several examples of policy
deployment show how the country used the catastrophe as a motivation factor to develop renewables.
Japan has overtaken Germany as the second biggest solar power generator in the world, only placing
behind mainland China. A mix of RPS (renewable portfolio standard), tax incentives to household solar
and mandatory country targets started the track, but a generous FIT compensation representing 171% of
the standard electricity tariff topped it off. These rates have fallen, mainly due to decline of solar PV price,
but also accounting for the adjustments due to the evolution of renewable energy generation costs. From
the set premium tariff, Japan has reduced to 120% in 2016 and Australia for example has fallen from
191%-313% to 21%-102% during the same period across its states [23]. These FIT premiums may
explain the large increase in the deployment of solar PV but also influencing the increase in FPV
installation.
Taiwan is another country that is trying to expand its installations by offering an 8.6% higher FIT for FPV
when compared to ground-mounted. With 200 lakes and ponds mapped as potential sites, meaning a
potential of 500 MW of solar energy, Taiwan expects FPV to support the achievement of the 20 GW solar
target by 2025 [24]. Dedicated tendering process are also a viable option, since it lets countries understand
the real price of a technology by receiving market offers, which turned to be the preferred policy scheme
in the past few years for renewables. Taiwan, China and the Indian state of Maharashtra, as well as several
others, have announced dedicated tendering for FPVs [16].
In the framework of policies, some considerations need to be addressed in order to develop a running
FPV market. Water bodies might require specific permitting and licensing between different types of
government agencies and industry. In the case of water treatment facilities, one has to secure that the
materials that are used in the facility are not hazardous to the water that will be sent for irrigation or
personal use, for example.
2.2.2 Panels
Until now, standard solar modules are used for FPV. One of the main concerns that can be raised is the
extended time the panels will be in moisture atmosphere. If the water contains salt or is near ocean shore,
the environment exposure will be rougher. However, man-made ponds usually consist of recycled or
drinkable water which do not contain a high level of salt like ocean waters. Moreover, reputable solar
panel manufacturers test their panels on certifications such as IEC 61701 (Salt Mist Corrosion Test) [25]
or ISO 9227:2017 (Corrosion tests in artificial atmospheres - Salt spray). The panels itself have a structure
that is robust against most typical environments found, as it can be seen in Figure 19 (a).
-31-
(a)
(b)
Figure 19: (a) Typical PV module assemble [26] (b) Polymeric frame PV [27]
The front glass of the module acts as the first protection, providing mechanical stability and rigidity
for the module. Encapsulates, normally made of EVA (ethylene vinyl acetate), are placed on top and at
the bottom of the solar cells in order to maintain a compact structure, giving adherence without damaging
the sensible cells and providing important feats such as protection against moisture and dirt penetration.
The back sheet is also made of a polymeric material that functions as an electric isolator and as an extra
protection against moisture penetration. Finally, the external frame of the panel is usually made of
anodized aluminum which is a highly corrosion-resistant metal [25], but could also be covered or
substituted by a polymeric frame (Figure 19 (b)).
2.2.2.1 Efficiency
Efficiency of Solar PV panels is normally presented in Standard Test Conditions (STC), which means that
the solar cells are tested with a normal irradiance of 1000 W/m², ambient temperature of 25 °C and 1.5
Air Mass atmosphere. As it is widely known, these conditions do not represent the normal conditions
found in the environment and do not account for several aspects that influence the conversion of light to
electricity. In the case of panels floating over water, some other aspects may induce higher or lower
efficiency levels.
The most important parameter that may increase the efficiency of FPV panels is the cooling effect of the
water. The performance has been measured in different ambient conditions, such as Portugal [18] and
South Korea [28][29] in hydropower dams and Brazil [30] in a microscale setup. All the measurements led
to increases in capacity factor between 7 to 10% when compared to ground-mounted systems.
Albedo, defined as the percentage of solar radiation that is reflected by a body compared to the total
incident irradiance, is believed to increase the overall efficiency by the high reflection of water surfaces.
However, it was observed on a testbed with 9 different floating structures that the albedo recorded was
only 5% and did not account to a substantial difference in energy output for traditional panels (as in
bifacial panels the situation may change) [31]. In fact, water albedo has been considered a complex
problem to model as it may depend on wavelength, wind intensity and characteristics of water surfaces,
and is not relevant when compared to the 20% ratio in grass and sand surfaces, 60% in snow [32] and
13% in rooftops [31].
-32-
2.2.3 Float structures
Today, there are several floating designs available in the market. Although the variety, they can be
classified in two types of structures: pure floats, most commonly used, or floats combined with metal
structures [16]. The first configuration consists of standard single buoyant bodies, normally produced with
HDPE (High Density Polyethylene) by blow molding. The second type of configuration may vary but
consists of metal structures or polymer structures, that serve as racks for the panels, that in turn are placed
over pontoons. Both layouts can be observed in Figure 20.
(a)
(b)
Figure 20: (a) Pure floats, adapted from [33] (b) Metal racks built over pontoons [32]
The main characteristic a pontoon should have is enough buoyancy to withstand the weight of the panels,
either individually, as the case of Figure 20 (a), or distributed in the case of Figure 20 (b). The footprint of
the buoys may influence the cooling effect and in a recent study, it was deemed statistically significant to
have less footprint and not so close to the water body [31]. Other characteristics may include resistance to
corrosion, UV protection and wind loads, but also accessibility to perform maintenance. Finally, and most
importantly for drink-water ponds, a certification of a regulatory institution enabling the use of the
materials over drinkable water.
2.2.4 Mooring and Anchoring
The mooring system is the one responsible to prevent free movement of the floating PV and to fix it in a
desired position, withstanding waves and wind loads. The difference between artificial ponds, such as
wastewater treatment or irrigation ponds, and offshore mooring for floating PVs is that the floating
structure is not exposed to high waves and tides. Moreover, if the location doesn’t have a relevant wind
speed on lower heights, the floating structure won’t be exposed to strong wind loads and the requirements
for the mooring design may be simpler than of those in marine environments or windy locations.
However, if not correctly sized it may not only be subdued to panel uplifting but also to a decrease in
energy output, as the whole structure may rotate towards a non-optimal direction (azimuth).
There is not a single solution for the mooring of floating PVs as many alternatives exist, including several
technologies utilized in the naval industry. Firstly, certain factors need to be considered when choosing the
mooring and anchoring. It must be designed for the extreme case which is when the pond is completely
full, which puts the highest strain on the ropes or chains connected to the anchor. As water levels
decrease, the cables will start to become loose which may cause the floating structure to move.
Bathymetry, that is the correspondent of topography but for water, is especially important to understand
the water profile and the depth. The type of water basin soil is also important, ruling if the anchoring, if
-33-
chosen to be on the bottom, will be able to penetrate the soil or not. In the case of wastewater treatment
ponds, where ponds are covered with a sealing liner, drilling the floor may cause permeation of
contaminated water to the groundwater. Anchoring to the bank is suitable for these cases and when the
ponds are small and shallow [16], which is especially useful in wineries irrigation ponds, for example, as
observed in the figures below.
Figure 21: (top) Rigid cabling mooring system anchored at the bank (bottom left) Rigid bracing (bottom right) and spring cables [17]
On the top of Figure 21 it is represented a rigid cabling setup fixed on the bank with concrete bollards,
which was probably sized for a pond that doesn’t have large variations in water level. Whereas in the
bottom part, it can be seen a more flexible setup, with a bracing that keeps the rigidity of the structure on
one side, while on the other there is a cable with springs that enable the system to respond to small
differences in water level.
Anchoring at the bottom of the water basin is also an option utilized by most of the floating installations
[16]. Naval technologies may be used here which includes catenary, taut leg mooring and other supportive
technologies. These setups are chosen normally for higher depths of water and may withstand higher
waves and winds. They may consist of cables, counter weights, auxiliary weights or floating bodies and a
concrete anchor. Figure 22 exemplifies some of the existing combination of types of mooring.
-34-
Figure 22: (left) Traditional dead-weight anchoring to the bottom [34] (right) auxiliary weight mooring coupled with pulleys [35]
Mooring and anchoring pose an engineering challenge to size and also affect costs depending on the level
of complexity and requirements. Although most of installations use traditional bank or bottom mooring
and anchoring, the Israeli company Solaris Synergy has developed a technology that breaks out of the
concept of fixed/semi-fixed mooring and promises grid parity and competitive costs compared to the
ground mounted installations. In Figure 23 it is possible to observe the grid proposed by the company,
which resembles a spider-web that is supported by a floating rim. The greatest advantage of this setup is
that it allows each panel to independently adjust to the wind loads, instead of the whole installation.
Figure 23: Solaris Synergy solution [36]
2.2.5 Cables and auxiliaries
Electricity and water should not mix, as this is a dangerous combination for humans and species that live
around the FPV. Therefore, as the power plants will be placed over water, two options are available to
transmit the current from the FPV to the inverter and transformers on land. One is by cabling it over
water, with the aid from buoys, and the second is by submerged cables. For both options, marine-graded
cabling is already used.
2.2.6 Tracking and concentrating
Depending on the design of the buoyant structure, floating PV also enables the inclusion of tracking and
concentration. Both systems will provide a better efficiency for the FPV overall and are useful when the
benchmark is to have the highest kWh/kWp installed.
-35-
Tracking for FPV have additional challenges, as the installation is not fixed and is subdued to fluctuations
of the water, which will interfere in the desired position of the PV panels. Two classifications for tracking
exist: within confining structure and without confining structure. The first is comprised of a structure
surrounding the PV arrays that is powered by an electric motor that rotates the whole structure towards
the desired position. Size is a limitation, since the bigger the power plant the larger diameter this structure
must have. For tracking without a confining structure, many solutions patented already exist such as using
ropes surrounding part of the structure connected to buoys with rollers and winches as well as submerged
structures moored to the center and connected to bow thrusters. Figure 24 below exemplifies two types of
confined and non-confined design, respectively.
(a)
(b)
Figure 24: (a) Confined tracking (b) Bow thrusters tracking [32]
Concentrating systems can also be used for FPV and can increase the energy yield per kWp installed.
However, it needs to be coupled with tracking and, sometimes, with additional cooling technology for the
panel, despite the cooling effect of the water. FPV with concentrating systems already exist (Figure 25)
and utilize low concentration setups, such as flat reflectors placed at rear of the module or in a V-shaped
setup with the module in the center. The first setup, may be able to increase energy yield up to 20%, but
does require a fully functional tracking [32] to be able to point the reflectors towards a perpendicular
orientation to the sun. This configuration also requires more area, which could be better used if another
panel could replace the reflectors and generate more electricity. The other option, the V-shaped
configuration, may add a larger concentration ratio but loses in energy yield as the panels, for simplicity,
will be set on a 0° tilt.
(a)
(b)
Figure 25: (a) Flat reflectors at the rear of neighboring panel (b) V-shaped concentration in Colignola, Italy [32]
-36-
2.2.7 Environmental impacts
Considering water is one of the most important elements in the world, installing a structure that is in
direct contact with it raises concern not only due safety but also if it will affect the environment. If the
water basin chosen for the FPV is to be used for irrigation or if it is drinkable water, one has to be extra
cautious about the requirements to utilize this surface, and, if they don’t exist, be prepared for it while
benchmarking other locations.
On a broader spectrum, installing FPVs in water basins may pose several direct impacts both in the
construction, operation and, if needed, decommissioning phase. Under construction, the main impact on
the water basin may occur due to the need of anchoring systems which may alter the geomorphology of
the soil. However, in certain projects such as from Ciel et Terre [33], the entire structure may be
commissioned outside of the water body and only if submersed mooring is required, a modification needs
to be done to the bottom of the basin. When considering the operation phase, these structures, together
with the floats or pontoons, will most likely develop algae which then may attract other species, in the case
of lakes, that will accommodate themselves below the FPV. It is not yet clear the impacts that this
behavior can cause, although in the case of some installations in Italy, it wasn’t noticed any dysfunction in
fauna nor flora due to the presence of the FPV [32]. For lakes with lively fauna, the covering of the
surface with PV panels may impact the luminosity and natural cycles of microorganisms which could be
negative to the ambient. For this reason, most of the commercially installed FPVs have been over artificial
ponds and sometimes abandoned mine pits that are harsh environments for the development of live
species.
One of the biggest issues, however, is understanding the compatibility of materials with water and how the
steady contact between metal structures and HDPE floats may affect water quality. Companies have been
working into certification of their materials in order to be able to install the structures over drinkable
water. Ciel et Terre has recently achieved the drinking water compliance with BS 6920:2000 of the Water
Regulation Advisory Scheme (WRAS), meaning that the materials are compatible with drinking water
exposure in the UK [33].
Considering the installations are rather new and haven’t even covered the lifetime of most of the
components, such as the panels, the mounting and inverters, further studies and analysis on existing
installations should be carried out in order to evaluate the relevance of the environmental impact of FPVs.
2.3 Wastewater treatment facilities
Human waste is a major water pollution media and has to be treated in order to be discharged back to the
environment. Examples of human waste are often from, but not limited to, urban cities’ residential sewage
system and manufacturing plants effluent discharge. It is composed by organic and inorganic materials
that were not initially present in such concentrations when the water was first consumed, therefore, it
requires treatment. Agricultural activity is also a major polluter by disposing unhealthy levels of drainage
containing agrochemical residues, as well as fertilizers which contain high levels of inorganic materials
such as nitrates and phosphates. Livestock may also contaminate water by the infiltration of liquid animal
waste in the soil reaching groundwater. The combination of such enriched nutrients can cause
eutrophication of the water body, leading to formation of algal blooms [37] and complicating the further
treatment. Such types of waste water may not only jeopardize the water quality, with the increased growth
of microorganisms, but also as a vector to transmission of pathogenic material such as cholera and
dysentery to humans.
Ergo, on a macro level, water treatment process should be able to remove suspended solids, ranging from
different sizes of particle, that not only decrease water quality for consumption but can also block the
pipes of the cities and even of the treatment facility. On a microsphere, it is desired to decrease the
biochemical oxygen demand (BOD), which means the ability of microorganisms to oxidize organic
-37-
material, remove inorganic material such as phosphates and nitrates and disinfect the water of fecal
coliforms and toxic material [37]. Further treatment, such as heavy metals removal, may also be required
when effluents come from manufacturing plants.
Depending on the final use of the water, wastewater treatment facilities may follow different requirements
to achieve the specification required, since the more operations are performed the costlier the final
treatment will be. In general, sewage treatment facilities will follow the ensuing routine [37]:
• Preliminary treatment: Removal of large solid materials that are thrown away by the population in
the sewage;
• Primary treatment: Normally consisted of sedimentation tanks, which will settle suspended solids
that have smaller particle sizes and couldn’t be removed by the previous process. This step,
together with the preliminary treatment, is important as the largest organic particles are removed
here, reducing BOD potential;
• Secondary treatment: Biologically operations, performed by heterotrophic bacteria that consume
organic material for its growth. Again, the main objective is to reduce BOD;
• Tertiary treatment: Biologically or chemically driven, used to remove the remaining organic ions.
More advanced tertiary treatments may be able to remove nitrates and phosphates, which are
relevant when combatting the overly rich nutrient water that can cause eutrophication. For these
steps, many advanced technologies can be used such as ozonation and reverse osmosis;
• Quaternary treatment: Focused on removing heavy metals and soluble minerals from water.
Although, compared to industrial effluents, sewage systems might not contain the same alarming
concentration of heavy metals, it might still be required depending on the regulation of the region.
Here, microalgae were found to be useful in the removal of heavy metals;
• Disinfection: Lay emphasis on preventing the spread of diseases by microorganisms such as virus,
bacteria and protozoa and destroying all possible pathogens. Many methods are used such as
boiling, ultraviolet irradiation and chlorine disinfection.
Although many different layouts exist to perform the treatment steps described above, this work will
focus on the waste stabilization ponds (WSP), because they enable the installation of a floating PV
structure.
2.3.1 Waste stabilization ponds method
WSP is a natural process method that is widely used for water treatment, which involves using
microorganisms such as bacteria to process organic and inorganic matters. Coupled with other
supplementary water treatment processes, stabilization ponds can be used to treat water up to tertiary
levels. These ponds are shallow basins, with large areas constructed and surrounded by land dikes that
isolate the ponds. Due to the high risk of seepage, which causes contamination of groundwater as well as
fluctuation in the water depth of the pond, sealing is required for these basins. Depending on the
application and requirements, several sealers can be used such as synthetic or rubber liners (Figure 26 (b)),
cement liners or chemical treatment sealers [22].
On sewage treatment facilities that use the WSP method, the following ponds can be found [37]:
• Anaerobic ponds: Usually 2-5 meters deep, receiving wastewater with heavy organic content.
Anaerobic bacteria are used to consume the organic waste which may generate methane and
particulate matter that is deposited at the bottom of the pond, hence the larger depth. Works at
neutral pH and retention times of up to 5 days;
• Facultative ponds: 1-2 meters deep and receive the wastewater coming from the aerobic ponds or
particle removing pre-processing. It is a mixture of anaerobic conditions, with sludge being
-38-
accumulated and digested by anaerobic bacteria at the bottom and aerobic conditions at the top,
with the mixing of oxygen and light for aerobic algae. Retention times goes up to 3 weeks;
• Maturation ponds: Aerobic digestion and last level of treatment to remove pathogens. Ponds
reach up to 1.5 meters of depth due to the need of the light to reach full depth. These ponds can
have aeration mechanisms in order to dissolve more oxygen to the bacteria, but the sun plays a
major role by increasing the temperature of the water, thus increasing the growth rate of bacteria
and dissolving more oxygen;
• Storage reservoir: Ponds utilized for the final storage of the clean water before being
commercialized or distributed to other tanks or holding ponds. Here the BOD level should be the
lowest [22] and algae and other microorganisms must be avoided.
As a continuous process, all the ponds operate at the same time and transfers are done considering the
retention time of each of them. Figure 26 (a) shows a typical wastewater treatment facility with several
ponds in series. Maturation ponds can be easily identified by observing heavy presence algae on the
surface. Ponds that have clear water could be either facultative ponds or water storage reservoirs.
(a)
(b)
Figure 26: (a) Ellis Creek Water Recycling Facility in Petaluma, California (retrieved from Google Maps) (b) Synthetic liner being laid out [22]
3 Case Study: Sonoma County, California
3.1 The US scenario
The United States have had a relation of ups and downs with sustainable energy affirmative action. From
not ever ratifying the Kyoto Protocol to its recent withdrawal from the Paris Agreement, the country has
put the globe on watch, since the country is one of the major polluters and consumers of energy in the
world. Such international agreements do not have any legal binding mechanism and, moreover, cannot
violate a country’s sovereignty and policies. Therefore, there is no guarantee that any of the proposed
targets will be reached by any of the ratified nations, as it depends largely on the will of people and
government action. Carbon emissions are so entrenched in any country’s economy and day-to-day lifestyle
that many factors have to be accounted when dealing with it. However, this doesn’t mean that the United
States haven’t taken any initiative against global warming. In fact, the electricity market and renewable
energy have benefited from several federal financial incentive policies enacted in the past decade, such as
the Production Tax Credit (PTC) and Investment Tax Credit (ITC). In 2016, 18.3% of total installed
power capacity and 15.6% of total annual electricity generation came from renewable energy sources,
respectively, meaning an increase of cumulative renewable installed capacity of 105% since 2006 [38].
-39-
3.2 The role of policies
At the same time, state legislation plays a major role in the United States. The 50 states, commonwealths
and territories, have their own legislation and courts to handle multiple issues, meaning they are also
sovereign and have the power to legislate everything that is not preempted by the federal constitution.
This might be one the reasons international agreements face so many difficulties when dealing with the
US, since every state has its own stakes. When analyzed under the scope of incentives for renewable
energy and efficiency, this becomes even clearer to observe, as big disparities exist between the states for
policies and targets. Figure 27 (a) exemplifies a scale and highlight some of the state’s numbers of policies
and incentives.
(a)
(b)
Figure 27: (a) State Incentives for Renewables & Efficiency adapted from DSIRE Database [39] (b) Renewable electricity installed capacity per capita adapted from [38]
Out of the most economic developed states, such as Texas, New York, Florida and Ohio, California
appears in the front-end of the scale with 229 policies focused on renewables and efficiency. On the
lower-end of the scale, states such as Wyoming, Kansas and North Dakota have almost two hundred less
policies focused on these matters. It doesn’t mean, however, that the state is not pushing enough towards
sustainability as in the same Figure 27 (b) it can be observed the highest spot for North Dakota in
renewable energy installed capacity per capita.
Nonetheless, not only California has the most policies enacted, but one of the most ambitious ones. In
September 2018, after subsequent amendments, the Renewable Portfolio Standard, which enforces electric
utilities to have a set share of the retail sales coming from eligible renewable energy resources, has again
increased the overall and gradual year requirements. Until 2030, 60% of the retail sales must come from
renewable energy and the target is to reach 100% by 2045. In comparison, at the same year of 2018, Green
Power Purchasing Goal for Federal Government was cut down from 30% by 2025 to 7.5% by the current
administration [39]. This shows that some states may move in the opposite direction of other states and
the federal government. In 2016, as a result of an ambitious and sustainable targets, California has reached
nearly 35 GW of renewable electricity capacity, followed by far from other states like Washington (25
GW) and Texas (22 GW) [38].
Out of the large renewable installed capacity, California also holds the number one spot in cumulative PV,
CSP and geothermal installations. Country wide, 25% of the PV installations are placed in California
(Figure 28), which is enforced by the several existing state incentives working along with the federal level,
as the mentioned Production and Investment Tax Credits. State policies include a set of financial
-40-
incentives and property tax exclusions for the state, but policies also exist on a county and municipal level.
Los Angeles, for example, holds its own net metering and Feed-in-Tariffs and San Francisco offers
financial incentives for multiple sectors that may reach up to US$ 50,000 for commercial installations. On
county levels, Sonoma County, offers a fixed interest rate below financial institute levels for solar related
projects [39]. Out of the few existing FPV installations in the US, California has shown an ideal location
for the development of projects. The first commercial FPV was installed in the Far Niente Winery, in
Napa Valley County and also one the largest projects in the US to date in Kelseyville, Lake County with
250 kWp DC [19] and in Dixon, Solano County with 607 kWp DC, only placing behind Sayreville’s, New
Jersey, 4.4 MWp DC.
Figure 28: 2017 PV installations by region in the USA [40]
3.3 The evaporation problem
The southwest region of the United States is known for its arid weather conditions, with the presence of
several strips of desert, such as the Mojave. The climate is very erratic and the dry seasons can be harsh,
leading to severe droughts felt by the population in the states of California, Nevada and Arizona.
Droughts are caused by less precipitation than expected and also less water coming from snow melting
from the mountains. El Niño and La Niña are also events that occur in the Pacific and influence the
weather in the western coast. Places where these conditions can become severe, have developed large
storage systems that help with the scarcity and higher demand of water. However, not only precipitation
and snow influence the aggravation of droughts, evaporation is also highly responsible, as water bodies are
out in the open, in hydropower dams and water storage ponds.
As nothing immediate can be done to force precipitation, California has had many initiatives to stop
evaporation from the existing water. The one that gained most notoriety was in 2015, when 96 million of
so-called “shade balls” were unloaded over a Los Angeles reservoir to initially prevent the formation of
bromate, a known carcinogen, instead of using chemical treatment. The HDPE balls also gained a known
second role which was to stop evaporation during dry seasons and it was reported that it could save up to
300 million of gallons, corresponding to the supply of 8,100 people. However, it was also reported that
the footprint of water consumption to produce these amount of plastic balls was up to 766 million gallons
[41], which meant that if locally produced would still provide a deficit of water in the first 2 years.
Evaporation is a process that occurs naturally without any harmful effect, but for certain regions in
specific seasons can become a severe problem. Wastewater treatment utilities are directly affected by
evaporation, especially in droughts, as water is essentially money. These industries spend energy by
pumping water and running mechanical machines, such as sludge scrapers in sedimentation tanks, as well
-41-
as in chemicals and biocides for treating water. Therefore, evaporation poses a direct treat in revenues.
Another major problem is irrigation, which is affected during droughts, as the water consumption must be
cut-back and groundwater may be used, which takes years to be replenished. Thus, having a structure that
prevents water evaporation is well desired for water bodies containing water to be commercialized.
3.4 Algal blooms
Suppressing the growth of unwanted algae in freshwater bodies is another benefit that is often related to
FPV. As the panels and floats provide shade to the water, it reduces the incoming radiation and, therefore,
reduces the capacity of any aerobic microorganism to perform photosynthesis and grow. It is not always
the case that algae and other microorganisms are unwanted, as it was shown in section 2.3.1, some ponds
in wastewater facilities utilize these microorganisms to help in the treatment process. However, once all
nutrients and waste are finally removed, the water that will be pumped to customers should abide to
specifications and not contain certain levels of bacteria or algae, for example.
There are several types of algae that may grow in freshwater bodies. On a visual inspection, a professional
may identify which type of algae is occurring, however only after microscopic analysis one may be able to
identify the exact specie. Common types of algae may be referred as Duckweed (known species: Lemna,
Spirodella), Watermeal (Wolffia) and Waterferns (Azolla). These are macroscopic plants that may cover
ponds in a few weeks during summer conditions, given the favorable environment for growth [42]. Since
algae are short-lived, it may result in high concentration of organic material which is then consumed,
inducing the growth of more microorganisms. This event is called Algal Bloom and can be seen in Figure
29 below.
(a)
(b)
Figure 29: (a) Duckweed in a pond [42] (b) Algal bloom seen from satellite in Lake Eire near Detroit [32]
Other common types of microorganisms and blooms may also occur in water bodies such as [42]:
• Floating cyanobacteria, referred as blue-green Algae (Microcystis aeruginosa, Aphanizomenon and
Anabaena): The bloom looks like a spilled oily green-blue paint and may be toxic to humans,
livestock and fish as it produces toxins that are dangerous for the liver;
• Floating or attached cyanobacteria, other types of blue-green algae (Oscillatoria, Lyngbya,
Phormidium and Plankthotrix): Found in floating mats or attached to the surface of the ponds;
• Attached cyanobacteria, blue green algae and known as “freshwater grapes” (Nostoc): Can also be
harmful to humans and livestock due to the production of liver toxins;
-42-
• Flagellated algae (Euglena, Phacus, Dinobryon, Prymnesium and Dinoflagellates): very strong odor
producing algae that can cause algal blooms in freshwater as well as in marine environments and
is very harmful to fish;
• Floating Diatoms: tend to form brownish films on the bottom or surface of the water basin.
These types of algae form the base of the food chain and may cause the growth of other
microorganism species in the water;
• Other algae such as filamentous (Spirogyra, Mougeotia and Zygnema) and branched filamentous green
algae (Cladophora and Hydrodictyon): thick filamentous algae that can be found attached to rocks.
Many ponds in the United States suffer from the unloading of water that contain high levels of nutrients
and set a perfect environment for algae to grow. It is not hard to find ponds that are suffering this type of
problem. Wainscott, New York, have had a high influx of nutrients being unloaded in the ponds, which
20% comes from wastewater and a third from agriculture [43]. Lake Eire (Figure 29 (b)) is another
location that year after year suffers from the appearance of algae blooms caused by runoff of household
water and farms. Wastewater facilities aren’t shielded from these problems either, as they have to treat
these types of wastewater and the processes aren’t 100% effective. Therefore, not only they might
contribute to the problem but might develop one of their own in their freshwater storage ponds.
3.5 Electricity prices
Wastewater treatment facilities are very energy intensive industries. In the US, these facilities may have an
electricity use intensity between the range of 0.375 kWh/m3 for larger utilities with treatment capacity
over a million of m3/day up to 1.12 kWh/m3 for smaller utilities with treatment capacity of 43,000 m3/day
[44]. Therefore, electricity is a major cost in these facilities, which may affect their budgets if large
fluctuations and increase in prices are not expected. In fact, average electricity prices in the US have
increased throughout the years, but California has showed a considerable larger increase in the past
decade. This can be observed in Figure 30, analyzing the industrial electricity prices over the span of
almost 10 years for the average of the United States and the state of California.
Figure 30: Average price of electricity to industrial consumers in the USA [45]
The reason for this increase, could be related to the high penetration of renewables in the energy mix, that
in the first moment need financial incentives, leading to higher prices. Since the RPS set is very high and
7.18 7.15 7.01 7.14 7.37 7.21 7.14 7.17 7.09
12.1211.02
12.21 12.2
13.94 13.76 13.6514.32 14.61
0
2
4
6
8
10
12
14
16
2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019
US$
cen
ts/k
Wh
Average price of electricity to Industrial consumers (base September)
Industrial USA Industrial California
-43-
the subsidies for renewables are attractive, this could also lead to phasing out of power plants, such is the
case of California, planning to retire three natural gas plants between October 2018 and September 2019
[45]. Ironically enough, floating photovoltaics could be the solution to help wastewater treatment
companies reduce their spending with electricity and producing clean and renewable energy.
3.6 Defining the location and resources
Given the above situation, where the need to adapt to new ambitious scenarios is an ongoing reality and
with the ever-increasing evaporation problems and electricity prices, it seems that California should be the
correct place to start the assessment. Out of all the states of the US, California is a leader in solar PV
installations, evidence to the high penetration of this technology and result of exclusive policies towards
the development of the solar in the industry sector. In order to select one of the several water bodies
available in this large state, a screening was done searching for wastewater companies already operating
with renewable energy and also with reported interest in installing solar PV. Many facilities already work
with solar energy, but it was found that one of the first wastewater treatment FPV in the US was placed in
a facility located in Kelseyville, Lake County, northern part of the state. The screening then proceeded to
investigate the region for other facilities, and it was found out that the Sonoma County Water Agency
(SCWA) installed a pilot FPV in one of its water storage ponds but has yet to install a larger power plant.
Therefore, since the agency operates several sites around the region with stabilization ponds and is open
for the implementation of this technology, it was decided to develop the thesis assessment within the
boundaries of a site operated by SCWA. Being a state-owned agency, almost all the information is made
public, which helps with the collection of modelling data. The site chosen, which can be seen in Figure 31,
is located between the cities of Windsor and Santa Rosa in Sonoma County and treats the water of the
Sonoma County Airport located just southwest.
Figure 31: Wastewater treatment site coordinates: 38°31'05.0"N 122°48'10.6"W (retrieved from Google Maps)
The location chosen refers to the Airport/Larkfield/Wikiup water treatment facility which is part of the
Airport/Larkfield/Wikiup Sanitation Zone (ALWSZ). The facility began operations in 1983 and covers
2,100 acres, treating water from approximately 3,850 Equivalent Single-Family Dwelling and designed to
provide treatment for 900,000 gallons of water per day (3,407 m3/day). The treatment level provided is
tertiary, meaning that the water should not contain organic ions, and is used for irrigation of grape and hay
fields in the region. On the left side of the figure, three aeration ponds are located side by side and next to
-44-
them two 95-million-gallon storage ponds (360 thousand m3) [46]. Another 100-million-gallon storage
pond, called Oceanview Reservoir, is also part of the facility and is located around 6 km to the west of the
main plant.
The region has interesting resource availability for PV. Using SAM database [47] for the coordinates of the
site shows high beam insolation with average levels of 6.06 kWh/m2/day, average temperature of 13.7 °C,
average wind speed at 10 meters height of 2.5 m/s and no presence of snow cover, which proves to be an
interesting site to place PV. In fact, on the figure it is possible to observe south of the olive-green pond,
three arrays of ground-mounted PV installed on the slope of the earth dikes. Therefore, the site shows
extreme potential for the operation of FPV by having large water body areas, high insolation, low wind
speed and openness of the agency that owns PV plants on the same location. Ultimately, the pond chosen
for the study will be the one located most north in the figure.
4 Modelling
Before the modelling starts, a second layout should be chosen to compare with the floating one. Most
urban settings will make the use of ground-mounted, rooftop and BIPV layouts, while offshore settings
could be used in coastal cities. A quick analysis over the Figure 31 shows that to match the potential of the
floating solar, the other layout to be compared should be the ground-mounted, as there is ground area that
is not being utilized. Moreover, there isn’t enough roof area and BIPV, as it has been previously
mentioned, is an alternative that is more related to aesthetics than to energy generation.
The modelling will start by feeding the location of the project in SAM in order to retrieve the irradiance,
ambient temperature, wind speed and dew temperature for the use in the models of PV, evaporation and
algae growth. Next, the output of these three models will be input for the LCOE and, together with the
results from the evaporation and algal growth models, a NPV analysis will be used to compare the two
projects economically. From the PV MATLAB model, other efficiency parameters will be compared such
as water temperature and capacity factor. The block diagram describing these steps can be seen in Figure
32.
Figure 32: Block diagram of the modelling
-45-
4.1 Design of the power plants
The most important design aspect of the proposed power plants is the energy demand in the wastewater
facility. For the financial year of 2018-2019, the agency is requesting a budged of 280,000 US$ [49]. If the
average price for electricity for the industrial sector in California is used, 14.61 US$ cents/kWh, the
average monthly consumption of the ALWSZ wastewater facility will be 159,708 kWh.
There are several ways to approach the sizing phase. One could be designing the solution to cover the
yearly demand of the facility, but to have the operation running fully on solar energy, batteries should be
installed, as these plants operate without interruption. From a benchmark of the market leader in the
development of FPV, it can be concluded that batteries haven’t been actively deployed yet [33]. For
ground-mounted PV, the capacity of distributed storage additions have increased in a higher proportion
than utility scale ones in the US between 2016 and 2017 [40], which may indicate that costs haven’t yet
reached an interesting range for larger projects.
The second approach could be installing as much capacity as the pond may fit. This would simplify the
analysis of evaporation and algae control, as the pond would be considered completely covered and the
models may represent a better picture of reality. However, with the area available reaching up to 70,700
m2, using a standard 350 W solar panel, the capacity could range from 4.8 MWp at the latitude tilt 38° to
more than 7 MWp at 12° tilt, commonly utilized by suppliers of FPV technology. These hypothetical
capacity could easily cover the demand of the facility and would either require batteries or electricity sold
to the grid by incentives like Feed-in-Tariffs. However, in California, FITs are only available through the
Renewable Market Adjusting Tariff (ReMAT) to capacities up to 3 MW [39]. Therefore, it has been
decided that the plants will be designed with 3 MWp capacity, covering the electricity demand of the
facility and eventually selling the surplus energy through the instrument available in the state while
respecting the limits to be eligible to the referred grant.
4.1.1 Common Design Parameters
When designing photovoltaics, some inputs are needed to be able to run a model, such as choosing a PV
panel, an inverter, the desired array size in power capacity, DC to AC ratio of the module/inverter,
tracking, orientation, tilt and losses associated to the DC and AC conversion. A very extensive database of
equipment can be found in SAM and multiple combinations of panels and inverters can be done, when
the limits of the chosen inverter are respected. Therefore, for simplicity, this project will use the panel and
the inverter found in a project located in Kelseyville (recall Figure 16 (a)) located in the same state, which
can be found with the full specification details on Appendix A and Appendix B.
The next parameter is the DC to AC ratio. There are many doubts that surround this parameter, since it is
advised as rule of thumb to be placed on unity, which means installing a 100 kW-DC array with a 100 kW-
AC inverter, for example. However, depending on the design goals the DC to AC ratio may change
slightly up to oversizing. The Swedish company ABB found out that regardless of site conditions, the ratio
to optimize specific yield (kWh/kW) should be placed below 1.2, while financial output objective sets this
ratio between 1.3 and 1.6 [50]. Since for this work, the comparison needs to be done considering the same
design parameters and the goal is to maximize energy yield, it was decided to set the DC to AC ratio to
unity. Oversizing could be performed to increase energy yield, but the model would penalize voltages that
are over the maximum DC voltage of the inverter, setting the string voltage to the maximum MPPT DC
voltage [51]. This behavior is called “inverter clipping” and is a safety mechanism to protect the
equipment against voltages over the nameplate capacity, which can be observed in Figure 33 below.
-46-
Figure 33: Inverter power clipping [50]
The tilt of the panels is also an important factor to be considered. Normally, the rule of thumb would tell
that the highest energy yield is achieved at the latitude angle. The tilt angle also affects the Ground
Coverage Ratio (GCR), which is the ratio of the module area to land area, meaning that a low number has
an interrow spacing that is large and a higher number means that this space is small. This is especially
important to check if the ground-mounted PV plant would be able to fit in the area that is available inside
the borders of the facility, using the optimal tilt of the panel. Therefore, considering a low Ground
Coverage Ratio (GCR) to avoid shading, a simulation was performed using SAM Parametrics function
varying the tilt and analyzing the annual output and specific yield, which are shown in Table 1.
Table 1: Energy yield (kWh) and specific yield (kWh/kW) for different tilt angles
Tilt (°) Annual energy
(kWh)
kWh/kW
0 1.64E+06 1475.54
20 1.87E+06 1687.9
25 1.90E+06 1712.9
30 1.91E+06 1726.83
31 1.92E+06 1728.3
32 1.92E+06 1729.32
33 1.92E+06 1729.91
34 1.92E+06 1730.06
35 1.92E+06 1729.76
36 1.92E+06 1729.01
37 1.91E+06 1727.81
38 1.91E+06 1726.17
39 1.91E+06 1724.07
40 1.91E+06 1721.53
-47-
Since area is not a constraint in the pond, as the power is limited to 3 MW, it is safe to say that the
selected tilt angle should be the one that gives the highest specific yield (kWh/kW). Therefore, 33° seems
the optimal angle when compared between the range from 0 to 40°, while also orientating the panels to
the south. Finally, tracking was not considered for neither of the plants, since no commercial tracking
technology has proven to be reliable for FPV yet; therefore, implementing a tracking system on the
GMPV would mean an unfair advantage for this case, as the kWh/kW would be improved for this
technology only.
With these parameters set, it is possible to simulate the generation of electricity per year and a capacity
factor. Most of SAM’s suggestions for default losses were considered on the custom model, which
accounts for 5% of soiling losses, that affects irradiance reaching the panels, 2% of module mismatch,
0.5% of diodes and connection losses, 2% of DC wiring and 1% of AC wiring losses. Regarding soiling
losses, it is argued that it is reduced for floating PVs as they sit in the middle of ponds away from shading
objects. However, no studies have confirmed this assumption, thus the same percentage used for the
ground-mounted PV will be applied to the FPV.
To calculate the final area required by the ground mounted PV, a set of trigonometric correlations were
used considering the orientation of the panels. To explaining these formulas, it will be considered here
that the panels are placed in a portrait layout, meaning that the shade is function of the larger side of the
panel (length L). The panels are to be placed with a surface azimuth angle (γs) of 180°, meaning that it
faces south, and a set tilt angle (β) equal to 33°. Considering the sun window between 9 a.m. and 4 p.m.
and using the Sun path chart available on Appendix C, the solar altitude angle (αs) is found to be 16.82°,
which gives the minimum distance in a diagonal (Diag) for the array as the equation below:
𝐷𝑖𝑎𝑔 = 𝐿 ∗sin(𝛽)
tan(𝛼𝑠) (𝐼)
However, the shading distance is given by correcting the diagonal distance when considering the solar
azimuth (γs) of 140.43°. Therefore, the horizontal distance d in the scheme of the Figure 34 is given by:
𝑑 = 𝐷𝑖𝑎𝑔 ∗ sin(180 − 𝛾𝑠) (𝐼𝐼)
Figure 34: Scheme representing the interrow spacing calculation
-48-
Finally, the interrow spacing between the strings of panels will be given by:
𝐼𝑛𝑡𝑒𝑟𝑟𝑜𝑤 = 𝑑 + 𝐿 ∗ cos(𝛽) (𝐼𝐼𝐼)
Therefore, knowing the interrow spacing and the number of modules that need to be installed, it is
possible to adequate a layout for the area available.
The main difference between a GMPV and a FPV lies in the model used for the cell temperature, which in
the case of the ground-mounted PV will be the NOCT model, commonly used in solar simulation
software, as seen below.
𝑇𝑐𝑒𝑙𝑙 = 𝑇𝑎𝑚𝑏 +𝐺
800∗ (𝑁𝑂𝐶𝑇 − 20) ∗ (1 −
𝜂𝑟𝑒𝑓
𝜏𝛼) ∗
9.5
5.7 + 3.8 ∗ 𝑉𝑤𝑖𝑛𝑑
(𝐼𝑉)
Where ηref is the reference module efficiency that is divided by the effective transmittance-absorbance
product (τα). Since SAM gives data for the wind speed at 10 meters height, the following correlation is
utilized to predict wind at one store or lower [51]:
𝑉𝑤𝑖𝑛𝑑 = 0.51 ∗ 𝑉𝑤𝑖𝑛𝑑@10 𝑚𝑒𝑡𝑒𝑟𝑠 (𝑉)
4.1.2 Floating PV design
For the floating PV design, the modelling to compute the energy yield will change slightly from the
ground-mounted one. The objective of this section is to show how it was performed through a set of
systems of linear and non-linear equations. For the calculations, the same panels, inverters, DC to AC
ratio, orientation, tilt and losses were applied.
4.1.2.1 Cell temperature model
As it was discussed during the literature review section, FPV have proven to be up to 10% more efficient
than the ground-mounted counterparts due to the cooling effect of the water just below the installation.
For the following sections, a cell temperature model that accounts for the water temperature will be used
based on the work of Kamuyu et al. [29] which models the cell temperature as follows:
𝑇𝑐𝑒𝑙𝑙 = −1.9034 + 1.12322 ∗ 𝑇𝑎𝑚𝑏 + 0.028655 ∗ 𝐺 − 0.6517 ∗ 𝑉𝑤𝑖𝑛𝑑 − 0.09362 ∗ 𝑇𝑤𝑎𝑡𝑒𝑟
− 0.001328 ∗ 𝑇𝑎𝑚𝑏2 − 0.000014 ∗ 𝐺2 + 0.08382 ∗ 𝑉𝑤𝑖𝑛𝑑
2 − 0.000604 ∗ 𝑇𝑎𝑚𝑏 ∗ 𝐺
− 0.031334 ∗ 𝑇𝑎𝑚𝑏 ∗ 𝑉𝑤𝑖𝑛𝑑 + 0.001389 ∗ 𝑇𝑎𝑚𝑏 ∗ 𝑇𝑤𝑎𝑡𝑒𝑟 − 0.000981 ∗ 𝐺 ∗ 𝑉𝑤𝑖𝑛𝑑
+ 0.000545 ∗ 𝐺 ∗ 𝑇𝑤𝑎𝑡𝑒𝑟 + 0.039145 ∗ 𝑉𝑤𝑖𝑛𝑑 ∗ 𝑇𝑤𝑎𝑡𝑒𝑟 (𝑉𝐼)
One of the limitations, however, is the absence of water temperature data of the chosen location. For that,
a prediction of water stream temperature will be used, based on Mohseni et al. [52], which modelled water
stream temperatures based on 584 US geological survey gaging stations utilizing non-linear regressions.
For the calculation of the water temperature, in function of air temperature, the following correlation was
used:
-49-
𝑇𝑤𝑎𝑡𝑒𝑟 = 𝜇 +𝛼 − 𝜇
1 + 𝑒𝛾∗(𝛽𝑎𝑖𝑟−𝑇𝑎𝑚𝑏) (𝑉𝐼𝐼)
Where α is the estimated maximum stream temperature, μ the estimated minimum stream temperature, γ
the steepest slope of the function and βair the air temperature at the inflection point. For the location’s
data, a visual inspection of the function demonstrates a S-shaped curve behavior (Figure 35), which for
the work of Mohseni et al. normally corresponded to a good fit with the Nash-Sutcliffe coefficient close to
unity, which represents the efficiency of the curve fit and should approach unity [52].
Figure 35: Correlation between air and water temperature
4.1.3 Single diode model
Although SAM is an open source software, some of the equation modules cannot be accessed yet, which
poses a difficulty for the simulation of a FPV using the same interface. Therefore, for the modelling of
floating solar, MATLAB [53] was used to build an algorithm following the guidance of SAM’s
Photovoltaic Model Technical Reference [51]. SAM uses the 5-parameter model, also known as the Single
diode model (Figure 36), as the equivalent circuit of a PV cell.
Figure 36: Single diode equivalent circuit of a PV cell [54]
-50-
The equivalent circuit starts with a light current source (Il), which stands for the photon-generated
current. The diode represents the diffusion of minority carriers in the depletion region of the p-n
junctions. In comparison to the ideal model, two resistances are added to the circuit which are the series
resistance (Rs) and the shunt or parallel resistance (Rsh), that represents the leakage current in the p-n
junction when the diode is in reverse biased region. The model is governed by the following equation:
𝐼 = 𝐼𝑙 − 𝐼𝑜 ∗ [𝑒𝑉+𝑅𝑠∗𝐼
𝑁𝑠∗𝑛∗𝑉𝑡 − 1] −
𝑉 + 𝑅𝑠 ∗ 𝐼
𝑅𝑠ℎ (𝑉𝐼𝐼𝐼)
The objective of this model is to retrieve the light current (Il), the reverse saturation current (Io), series
resistance (Rs), shunt resistance (Rsh) and the diode ideality factor (n). However, since for the panels
chosen, all the parameters on the reference conditions are already given, the parameters will only be
function of either temperature, irradiance or both. SAM designates an ideality factor (a) that represents the
product of the number of cells in series (Ns), diode ideality factor (n) and thermal voltage (Vt) (the
denominator inside the exponential of the Equation VIII).
For a set of weather hourly data extracted from SAM database, which consists of calculated irradiance
after soiling losses, ambient temperature and wind velocity, vectors of the parameters were calculated
following a sequential procedure. The light current varies with the irradiance (G) and temperature of the
cell (Tcell) given by:
𝐼𝑙 =𝐺
1000∗ (𝐼𝑙𝑟𝑒𝑓 + 𝛼𝑖 ∗ (𝑇𝑐𝑒𝑙𝑙 − 298.15)) (𝐼𝑋)
Where αi stands for short circuit current temperature coefficient, that was considered constant. The
reference conditions light current (Ilref) was retrieved from the chosen module. For the reverse saturation
current, the reference conditions (Ioref) were also given and, for a set of conditions, this current is
dependent on the cell temperature as follows:
𝐼𝑜 = 𝐼𝑜𝑟𝑒𝑓 ∗ (𝑇𝑐𝑒𝑙𝑙
295.15)
3
∗ 𝑒(
1𝑘
)∗(1.12
𝑇𝑐𝑒𝑙𝑙𝑟𝑒𝑓−
𝐸𝑏𝑔𝑇𝑐𝑒𝑙𝑙
) (𝑋)
𝐸𝑏𝑔 = 1.12 ∗ (1 − 0.0002677 ∗ (𝑇𝑐𝑒𝑙𝑙 − 𝑇𝑐𝑒𝑙𝑙𝑟𝑒𝑓) (𝑋𝐼)
To solve Equation (VIII) one needs to find the open circuit voltage (Voc) and short circuit current (Isc)
for each of the weather conditions. This means in practical terms, building an IV-curve for each hour of
weather data that is collected. At the open circuit voltage, the current equals zero and the resulting
equation is as follows:
𝑉𝑜𝑐 = 𝑅𝑠ℎ ∗ ( 𝐼𝑙 − 𝐼𝑜 ∗ [𝑒𝑉𝑜𝑐
𝑎 − 1]) (𝑋𝐼𝐼)
The ideality factor and the shunt resistance vary with the temperature and irradiance, respectively below:
-51-
𝑎 =𝑇𝑐𝑒𝑙𝑙
𝑇𝑐𝑒𝑙𝑙𝑟𝑒𝑓∗ 𝑎𝑟𝑒𝑓 (𝑋𝐼𝐼𝐼)
𝑅𝑠ℎ =1000
𝐺∗ 𝑅𝑠ℎ𝑟𝑒𝑓 (𝑋𝐼𝑉)
To find Voc, Equation XII must be solved numerically which is possible by using the function “fsolve” in
MATLAB. For the short circuit current, the voltage is set to zero and the series resistance is considered to
be constant, giving the following correlation which also has to be solved numerically.
𝐼𝑠𝑐 = 𝐼𝑙 − 𝐼𝑜 ∗ [𝑒𝑅𝑠∗𝐼𝑠𝑐
𝑎 − 1] −
𝑅𝑠 ∗ 𝐼𝑠𝑐
𝑅𝑠ℎ (𝑋𝑉)
To calculate the DC power output of the designed plant, the MPPT is considered as the operating on the
IV-curve given by the weather conditions. Therefore, for every hour of data it is needed to calculate the
Vmp and Imp to compute the energy being produced by one single module. The equations that need to
be solved to obtain both parameters are rather complicated and form a system of two non-linear
equations. For this work, the solution used was based on the work of Hejri et al. [54] that computes the
following equations:
𝐼𝑚𝑝 = −𝑉𝑚𝑝 ∗
((−1
𝑅𝑠ℎ) −
𝑅𝑠ℎ ∗ 𝐼𝑠𝑐 − 𝑉𝑜𝑐 + 𝑅𝑠 ∗ 𝐼𝑠𝑐𝑎 ∗ 𝑅𝑠ℎ
∗ 𝑒𝑉𝑚𝑝+𝑅𝑠∗𝐼𝑚𝑝−𝑉𝑜𝑐
𝑎 )
(1 +𝑅𝑠
𝑅𝑠ℎ+ 𝑅𝑠 ∗ (
𝑅𝑠ℎ ∗ 𝐼𝑠𝑐 − 𝑉𝑜𝑐 + 𝑅𝑠 ∗ 𝐼𝑠𝑐𝑎 ∗ 𝑅𝑠ℎ
∗ 𝑒𝑉𝑚𝑝+𝑅𝑠∗𝐼𝑚𝑝−𝑉𝑜𝑐
𝑎 ))
(𝑋𝑉𝐼)
𝐼𝑚𝑝 = 𝐼𝑠𝑐 − (𝐼𝑠𝑐 −𝑉𝑜𝑐 − 𝑅𝑠 ∗ 𝐼𝑠𝑐
𝑅𝑠ℎ) ∗ 𝑒
𝑉𝑚𝑝+𝑅𝑠∗𝐼𝑚𝑝−𝑉𝑜𝑐𝑎
−𝑉𝑚𝑝 + 𝑅𝑠 ∗ 𝐼𝑚𝑝 − 𝑅𝑠 ∗ 𝐼𝑠𝑐
𝑅𝑠ℎ (𝑋𝑉𝐼𝐼)
Once the maximum power point is calculated, the maximum DC power is computed as follows:
𝑃𝑑𝑐 = (𝑉𝑚𝑝 ∗ 𝐼𝑚𝑝) ∗ 𝐷𝐶𝐿𝑜𝑠𝑠𝑒𝑠 (𝑋𝑉𝐼𝐼𝐼)
Next, the DC power should be converted into AC power. For that, first it is checked how many of the
selected modules in series the inverter can withstand without reaching the maximum DC voltage. Since
the modules open circuit voltage at reference condition is 46.6 V and the inverter maximum DC voltage is
800 V, only 17 strings are allowed in series per inverter. The next step is to compute the string voltage
operating at the maximum power point found by the two non-linear equations. The proposed model caps
the voltage between the maximum and the minimum MPPT DC voltage of the inverter, which means that
it penalizes string voltages above 800 V and improve voltages that are below 600 V. Ergo, having the
actual string voltages, the power per string can be calculated as follows:
𝑃𝑎𝑐𝑠𝑡𝑟𝑖𝑛𝑔 = 𝑃𝑑𝑐𝑠𝑡𝑟𝑖𝑛𝑔 ∗ 𝐴𝐶𝐿𝑜𝑠𝑠𝑒𝑠 ∗ 𝜂𝑖𝑛𝑣𝑒𝑟𝑡𝑒𝑟 (𝑋𝐼𝑋)
-52-
The number of strings, panels and installed power required to top the demand of the wastewater
treatment facility will be given below. Additionally, the capacity factor (CF) of the power plant is
calculated to be able to compare with the ground-mounted one and other projects.
𝑆𝑡𝑟𝑖𝑛𝑔𝑠 =𝐸𝑛𝑒𝑟𝑔𝑦 𝐷𝑒𝑚𝑎𝑛𝑑
𝑃𝑎𝑐𝑠𝑡𝑟𝑖𝑛𝑔 (𝑋𝑋)
𝑁𝑝𝑎𝑛𝑒𝑙𝑠 = 𝑃𝑎𝑛𝑒𝑙𝑠 𝑝𝑒𝑟 𝑠𝑡𝑟𝑖𝑛𝑔 ∗ 𝑆𝑡𝑟𝑖𝑛𝑔𝑠 (𝑋𝑋𝐼)
𝐷𝐶𝑝𝑜𝑤𝑒𝑟 = 𝑁𝑝𝑎𝑛𝑒𝑙𝑠 ∗ 𝑃𝑚𝑝𝑟𝑒𝑓 (𝑋𝑋𝐼𝐼)
𝐶𝐹 =𝐸𝑛𝑒𝑟𝑔𝑦 𝐷𝑒𝑚𝑎𝑛𝑑
24 ∗ 365 ∗ 𝐷𝐶𝑝𝑜𝑤𝑒𝑟 (𝑋𝑋𝐼𝐼𝐼)
5 Economic Analysis
In this section, the economic feasibility will be analyzed through the scope of the levelized cost of
electricity (LCOE) and Net Present Value (NPV). At first, the LCOE will be evaluated for the floating
photovoltaic and compared to the proposed GMPV and well documented industry values. Several
information will be needed to proceed with this evaluation, mainly the investment costs of FPV and
maintenance costs, that remain unclear due to the reduced number, maturity and track record of the
projects. Moreover, the industry is highly concentrated on a few players, making it harder to obtain new
and different standpoints. Next, the NPV of the two proposed configurations will be analyzed,
considering the different types of revenues that may exist and the variable costs. Three LCOE and NPV
scenarios will be simulated as follows:
• Comparison between the ground-mounted (GMPV) and FPV with and without ITC, not
considering extra benefits of FPV. This scenario will be considered the Business-as-Usual (BAU)
scenario;
• Comparison between both projects with and without ITC, considering possible benefits from
evaporation and algal suppression. This scenario will be considered the Extra benefits scenario;
• Comparison between both projects with and without ITC, considering the acquisition of extra
land for the ground-mounted layout. This scenario will compare the FPV of the above two
scenarios against a GMPV requiring extra land;
Therefore, covering these three scenarios, one may identify which category to frame its one’s reality and
verify which project is then the most profitable.
5.1 Levelized Cost of Electricity
The levelized cost of is a very commonly used metric, which analyzes the ratio of the lifecycle cost and the
lifetime energy production of a project given by the formula below.
𝐿𝐶𝑂𝐸 =
∑It + 𝑂𝑀𝑡 + 𝐹𝑡
(1 + 𝑟)𝑡
𝑛
𝑡=1
∑𝐸𝑡
(1 + 𝑟)𝑡
𝑛
𝑡=1
(𝑋𝑋𝐼𝑉)
-53-
Where It stands for the initial investment costs that may start before the year 1, OMt stands for Operation
& Maintenance costs, Ft stands for the fuel costs and Et stands for the electricity generation. All these
parameters are evaluated along their lifetimes and have to be adjusted by a discount rate. Using the time
value of money, this rate is used to bring a future value to the present to be able to represent a current
situation. The rate is normally fixed, as it is hard to predict how the market and interest rate change with
time. Increase in capacity, salvage values and tax incentives may also be included in the formula and will
always be evaluated within the discount rate.
For floating PV, the costs are still unclear, as several technologies still find their way into the market, few
projects exist, when compared to ground-mounted, and are limited to certain geographies, mainly Japan
and China. This has to do with the fact that several types of mounting exist and anchoring still plays a
major role in costs as it has been previously mentioned in the above sections. Through extensive research
from media releases and industry information it was possible to summarize in Figure 37, Figure 38 and
Figure 39 different FPV project costs by capacity installed. The vertical axis stands for the investment cost
per Wp installed, the size of the bubble represents the size of the power plant (the size legend for each
plot is on the upper right corner) and the color of the bubble represents the region. The details of the
power plants represented here can be found in Appendix D.
Figure 37: Specific cost of Floating PV projects below 3 MW (Compilation based on media releases and industry information on Appendix D)
As with every new technology, the power plants started with capacities below 500 kW and spread around
the world, with the first two being built in the US, but with Japan soon taking over on the deployment of
most of the projects in the first decade since the inception of FPV. Throughout time, it is possible to
observe the decrease in costs to the range of 2 to 3 US$/Wp, with a slight advantage for installations in
Europe.
0
2
4
6
8
10
12
2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020
Inst
alla
tio
n c
ost
s (U
S$/W
p)
Floating PV cost per Wp in the world (Power < 3MW)
Australia Belgium Brazil Chile Colombia
India Italy Japan Malaysia Netherlands
Portugal South Korea Spain UK USA
3 MW
2 MW
1 MW
0,5 MW
Size Legend
-54-
Figure 38: Specific cost of Floating PV projects between 3 MW and 10 MW (Compilation based on media releases and industry information on Appendix D)
As FPV gained maturity, more projects started to appear with capacities over 3 MW, but still lacking a
relevant improvement in costs, as it still sit around 2 US$/Wp.
Figure 39: Specific cost of Floating PV projects over 10 MW (Compilation based on media releases and industry information on Appendix D)
Finally, as larger scale projects started to appear, mostly deployed in China, the price per Wp reduced
significantly with time. Japan holds the majority of the floating PV plants but still has a higher price per
0
1
2
3
4
5
6
7
2014 2015 2016 2017 2018 2019 2020
Inst
alla
tio
n c
ost
s (U
S$/W
p)
Floating PV cost per Wp in the world (3 < Power < 10 MW)
China Brazil Japan South Korea UK USA
8 MW
5 MW
3 MW
Size Legend
0
0.5
1
1.5
2
2.5
2015 2016 2017 2018 2019 2020
Inst
alla
tio
n c
ost
s (U
S$/W
p
Floating PV cost per Wp in the world (>10 MW)
China Japan Vietnam France South Korea
150 MW
100 MW
40 MW
20 MW
Size Legend
-55-
watt, whereas China has started to connect new power plants over 100 MW and Indian tenders for power
plants of the same size have been pushing the costs below 1 US$/Wp [16].
The United States has just a few active projects, but several municipalities have been studying the
implementation of FPV. Costs have come down as far as 1.6 US$/Wp for the largest project in the US in
Sayreville, New Jersey, but maintained around 3-5 US$/Wp for plants below 1MW as can be seen in
Figure 40.
Figure 40: Specific cost of Floating PV projects in the USA (Compilation based on media releases and industry information on Appendix D)
When compared to the benchmark data, commercial ground-mounted PV of 100 kW and 1 MW have
been evaluated between the range around 1.95 to 1.72 US$/Wp, respectively, as California appears at 1.92
US$/Wp for a standard 200 kWp, meaning that however the good availability of resources, it is still one of
the most expensive states. Fixed tilt utility scale range between 1.36 to 1.06 US$/Wp for 5 to 100 MW,
respectively [55]. Therefore, as expected with new technologies, implementation costs are relatively higher
and become more competitive once large-scale projects are deployed. For this work, a research on the
several components included in ground mounted and FPV were researched [55], [56] and summarized in
Table 2 below.
0
1
2
3
4
5
6
7
8
9
10
11
2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019
Inst
alla
tio
n c
ost
s (U
S$/W
p
Floating PV cost per Wp in the USA
New Jersey Florida California Colorado
4 MW
0,5 MW
Size Legend
-56-
Table 2: Cost breakdown of ground-mounted and floating PV
Ground-Mounted
PV
Floating PV
Size 3 MWp 3 MWp
Panels (US$/Wp) 0.350 0.350
Inverter (US$/Wp) 0.180 0.180
Structural components (US$/Wp) 0.210 0.690
Electrical components (US$/Wp) 0.250 0.250
Transmission line (US$/Wp) 0.005 0.005
Installation labor (US$/Wp) 0.004 0.001
EPC overhead 10% over equipment costs and 40%
over installation labor
Burden rates 31.8% over Installation labor
Contingency 3% mark-up on EPC cost
PII (US$/Wp) 0.050 0.050
Developer’s Profit 8%
Sales Tax 8.125%
Fixed Costs after sales tax (US$/Wp) 1.384 2.007
For the elaboration of this table, the PV plants were categorized under the commercial to large scale
projects, which means that it may benefit from reduction of costs by scale. Under structural components,
ground-mounted PVs consider rack mounting, whereas FPV costs account for the floats, screws, rivets
and elastic joints to keep the structures cohesive as well as the foundation and anchoring. For the type of
float, it was decided to choose the modular ones made of HDPE material (recall Figure 20 (a)), which
provide a better flexibility to assembly and is the structure most of the projects are based on. Electrical
components encompass the conductors, conduits, fitting boxes, panel boards and all other components
that both share in common, like transmission lines. Over these components costs, it was considered an
EPC (Engineering, Procurement and Construction) overhead cost of 10%, which is between the normal
practice from the market range of 8 and 13%, and a contingency mark-up of 3% over these same
components [55]. Over the overall installation costs, solar projects are eligible to the Federal Investment
Tax Credit (ITC) of 30% if the construction starts before the end of 2019 and reduces gradually to 26% in
2020 and 22% in 2021. For this project, it will be considered to be under construction in the year of 2020,
therefore the subsidy of 26% will be applied [39].
Labor costs were considered 40 US$/h for the ground-mounted PV, including electricians and laborers,
while for the floating PV it was increased to 60 US$/h as depending on the mooring it is necessary
perform work under water, which could potentially increase costs. Although the labor costs are more
expensive per hour, the installation time for a ground-mounted project was considered be 8 weeks due to
the need to prepare the land and build the foundation, while FPV with modular floats was considered to
take up to a week. Over labor costs, 40% overhead cost due to unexpected events was applied as well as
burden rates that are related to workers compensation (31.8% of the direct labor cost).
-57-
Moreover, the following costs were also considered: permitting, inspection, and interconnection (PPI), a
5% profit for the developer and Windsor’s, California, sales tax of 8.125% which applies over the
good and services provided in this project.
Operation and Maintenance (O&M) costs haven’t been well documented for FPV, but the maintenance
procedures follow the same as from ground-mounted. The on-site availability of water may be an
advantage in the cleaning process of floating PVs when compared to ground-mounted ones, but the
presence of mooring lines and anchors can also mean that the O&M can be higher, due to the need for
inspection underwater. For ground-mounted PV it was considered a cost of 0.013 US$/MW/year [55],
while for FPV it was considered 0.026 US$/MW/year.
The last term of the LCOE equation is the energy generated through the lifetime of the project. Lifetime
of the project was considered to be 25 years, in which the performance degrades 0.5% per year, starting to
occur after year 2 [51]. Finally, the discount rate was considered to be 6.75%, taken from a recent survey
performed by the professional services network Grant Thornton, where hundreds of investors,
representing billions of pounds around the world, answered what was the closest match for a discount rate
for renewables projects such as solar, wind and hydro [57].
5.2 Net Present Value
The NPV represents all the future cash flows, both positive and negative, i.e. revenues and maintenance
costs, over the lifetime of the power plant discounted to the present. It represents the profitability of a
project and is used to compare between alternative projects which one to prioritize. The NPV can be
calculated by the following equation:
𝑁𝑃𝑉 = − ∑𝐶0𝑡
(1 + 𝑟)𝑛
𝑛−1
𝑡=−𝑁
+ ∑𝐶𝑡
(1 + 𝑟)𝑛
𝑛
𝑡=1
(𝑋𝑋𝑉)
Where C0 represents the initial investment, which may be considered being done at the year 0 of the
project which then simplifies that equation to a -C0 term. It also may start N years prior to year 0 and
therefore should be discounted to the future. Ct represents the discounted revenues that the project will
earn during its lifetime. In the case of both PV, the surplus of the energy generated will be sold to the grid.
In this region, the grid is managed by Pacific Gas & Electric Company (PG&E) that, through the ReMAT
tariff, set the selling prices at US$ 89.23/MWh for baseload and as-available non-peaking resources and
US$ 61.23/MWh for as-available peaking resource which is typically where solar PV is framed. The
delivery term of such grant can be chosen for 10, 15 or 20 years [39].
The NPV will be calculated for both power plants and a new scenario will be analyzed which accounts for
the extra benefits due to the presence of the FPV. In this scenario, the increased revenue of water that has
not been evaporated and the reduction of the maintenance costs of algae control will be credited as
revenues to the NPV of the FPV.
5.3 Evaporation
One of the most hyped extra benefits from the FPV is the evaporation reduction. Evaporation is the
physical process in which liquid water converts to water vapor coming from surfaces as soil, water bodies
and vegetation. Evapotranspiration (ET) is another process commonly referred for the conversion of
water in wet vegetation and crops. Natural evaporation is a hard parameter to measure, especially for large
-58-
water bodies and several authors have presented models that try to study this phenomenon. These
processes depend to certain extent on wind speed, solar radiation, air temperature and the type of soil.
For the evaporation model, the equation of Penman-Monteith [58] was chosen due to its reference in the
hydrological studies. The equation can be seen below:
𝐸𝑇𝑜 =0.408 ∗ (𝑅𝑁 − 𝐺𝑠𝑜𝑖𝑙) + 𝛾 ∗
900𝑇𝑎𝑖𝑟 + 273
∗ 𝑢2 ∗ (𝑒𝑠 − 𝑒𝑎)
𝛥 + 𝛾𝑝 ∗ (1 + 0.34 ∗ 𝑢2) (𝑋𝑋𝑉𝐼)
Where ETo stands for the reference evapotranspiration rate (mm/day), Tair is the mean air temperature
(°C), u2 is the wind speed at 2 meters height (m/s), Rn is the net radiation flux (MJ/m2day), Gsoil is the soil
heat flux (MJ/m2day) that is considered zero for water, Δ is the slope of the vapor pressure-temperature
curve (kPa/°C ), γp is the psychrometric constant (kPa/°C ), es is saturated vapor pressure (kPa), and ea is
actual vapor pressure (kPa).
The net radiation flux (Rn) is a parameter that is hardly measured and the database is very scarce. The
physical meaning of this parameter is the balance between incoming and outgoing energy, which rules the
interaction between soil and atmosphere and can influence the climate. To be able to compute the
Penman-Monteith equation without measuring the net radiation flux, empirical equations can be used to
determine this parameter. The equation below is proposed to compute this flux [59]:
𝑅𝑛 = (1 − 𝛼𝑎𝑙𝑏) ∗ 𝑅𝑠𝑖 − (1.35 ∗ (𝑅𝑠𝑖
𝑅𝑠𝑜) − 0.35) ∗ (0.35 − 0.14 ∗ 𝑒𝑎
0.5) ∗ 𝜎 ∗ (𝑇𝑚𝑎𝑥
4 − 𝑇𝑚𝑖𝑛4
2) (𝑋𝑋𝑉𝐼𝐼)
Where αalb represents the albedo, σ the Stefan-Boltzmann constant (5.67x10-8 W/m2K4), Rsi is the
measured solar radiation (W/m²), Rso represents the clear sky solar radiation. Tmax and Tmin are the
maximum and minimum daily temperature measured. To compute the parameter of the net radiation flux
the following equations are used:
𝑒𝑎 = 0.611 ∗ exp (17.27 ∗𝑇𝑑𝑒𝑤
𝑇𝑑𝑒𝑤 + 237.3) (𝑋𝑋𝑉𝐼𝐼𝐼)
𝑅𝑠𝑜 = (0.75 + 0.00002 ∗ 𝐴𝑙𝑡𝑖𝑡𝑢𝑑𝑒) ∗ 𝑅𝑠𝑎 (𝑋𝑋𝐼𝑋)
Where Rsa represents extra-terrestrial solar radiation expressed as a function of solar constant (0.08202
MJ/m2min), the relative distance between Earth and sun (dr), sunset time angle (ωs), latitude (φ), solar
declination (δ) and day of the year (J).
𝑅𝑠𝑎 = (24 ∗ 60
𝜋) ∗ 𝐺𝑠𝑐 ∗ 𝑑𝑟 ∗ (cos(𝜑) ∗ cos(𝛿) ∗ sin(𝜔𝑠) + 𝜔𝑠 ∗ sin(𝜑) ∗ sin(𝛿)) (𝑋𝑋𝑋)
𝑑𝑟 = 1 + 0.033 ∗ cos (2𝜋𝐽
365) (𝑋𝑋𝑋𝐼)
-59-
𝜔𝑠 = cos−1(−tan (𝜑) ∗ tan(𝛿)) (𝑋𝑋𝑋𝐼𝐼)
With these parameters in hand, physical parameters that comprise the Penman-Monteith equation are
calculated, where heat of vaporization of water λvap (MJ/kg) is given as:
𝜆𝑣𝑎𝑝 = 2.501 − 2.361 ∗ 10−3 ∗ 𝑇𝑎𝑚𝑏 (𝑋𝑋𝑋𝐼𝐼𝐼)
𝛾𝑝 = 𝑃 ∗𝐶𝑝𝑎𝑖𝑟
0.622 ∗ 𝜆𝑣𝑎𝑝 (𝑋𝑋𝑋𝐼𝑉)
𝑒𝑠 = 0.6108 ∗ exp (17.27 ∗ 𝑇
𝑇 + 237.3) (𝑋𝑋𝑋𝑉)
𝛥 =4098 ∗ (0.6108 ∗ exp (
17.27 ∗ 𝑇𝑇 + 237.3
))
(𝑇 + 273.2)2 (𝑋𝑋𝑋𝑉𝐼)
Finally, SAM records the wind speed (m/s) at 10 meters height, so to use it in the equation it needs to be
converted to 2 meters height following the conversion below:
𝑢2 = 𝑢ℎ ∗4.87
ln(67.8 ∗ ℎ − 5.42) (𝑋𝑋𝑋𝑉𝐼𝐼)
Substituting all the values on the Penman-Monteith equation, one can find the yearly evaporation rates in
mm/day for a given location. While placing the PV panels over a percentage of the area of the pond, the
evaporation is expected to be reduced since it is function of the area exposed to the influencing
parameters. Therefore, the difference between a normal condition without cover and another scenario
with the FPV will give the water that could be saved from evaporation. For this model, when covering the
surface with shade, the evaporation is considered eliminated. The economic benefit of a reduction in
evaporation will come from extra sales of water. The price of water for the Sonoma County Water Agency
has more than doubled in the last few years as it can be observed in the Figure 41 for all the aqueducts
operated. This could mean that overall operating costs have also increased, so saving on water evaporation
will give a larger return to the Agency.
-60-
Figure 41: Historical water rates profile per aqueduct [60]
5.4 Algae growth
Although highlighted as a potential benefit of FPV, algae growth suppression for this application hasn’t
yet been evaluated and published in relevant literature. There are hundred thousand of algae species that
could develop in water and for every type of species, a photosynthesis and growth model may apply.
These models should be coupled with experimental analysis but are most limited to laboratory
experiments and have not been validated in outdoor conditions yet. Therefore, finding a model that suits
the growth rate of the potential algae inhabiting the water body in the given location is a challenging task.
Many models were researched, screening for the ones that looked on common algae that may occur in
freshwater, briefly described in Section 3.4, and that could withstand the ambient temperature conditions.
To approach this problem, it was decided to consider light and temperature acting on the growth rate of
algae differently, as independent factors. The mathematical model proposed by Talbot et al. [61] tries to
outline relationships for common parameters related to growth rate such as light intensity for species of
algae that are common occurrences in freshwater such as Oscillatoria agardhii and Phormidium bohneri and will
be used to retrieve the parameters required for the calculations. The model proposed by Bernard &
Rémond [62] will be used to calculate the growth rate expected by covering the pond. Therefore, this
section will display the equations that will be used to exemplify the effect of light shading in the growth
rate of these types of algae.
The first step is to convert the daily irradiance data given in W/m² to quantum flux of light (QF) which is
the measure most important for biological processes, given in μmol/m²s. For that, the following
correlation is used [63]:
𝑄𝐹 = 𝐺 ∗ 𝜆 ∗ 0.00836 (𝑋𝑋𝑋𝑉𝐼𝐼𝐼)
Plants and algae have their singular and special sensitivity for photosynthesis and absorb different
wavelength of the light spectrum (λ), which is closely related to the most abundant pigment present in the
algae structure. For cyanobacteria, such as the species analyzed, the most common pigments found are
Chlorophyll-a, phycocyanin and phycoerythrin which each have a preferential wavelength of absorption.
0.00
0.10
0.20
0.30
0.40
0.50
0.60
0.70
0.80
0.90
2000 2005 2010 2015 2020
Rat
e (U
S$/m
³)
Water rates per aqueduct
Santa Rosa Aqueduct Petaluma Aqueduct Sonoma Aqueduct
-61-
For simplicity, the wavelength absorbed will be considered 600 nm, which is the last two pigments’
preferred wavelength [64].
Next, the growth rate of the studied algae (μg) is given by [62]:
𝜇𝑔 = 𝜇𝑜𝑝𝑡(𝑄𝐹) ∗ 𝛷(𝑇) (𝑋𝑋𝑋𝐼𝑋)
Where the optimal growth rate (μopt) and the parameter Φ are given by:
𝜇𝑜𝑝𝑡 = 𝜇𝑚 ∗𝑄𝐹
𝑄𝐹 +𝜇𝑚𝛼𝑔
∗ (𝑖)2
𝛷 =(𝑇 − 𝑇𝑚𝑎𝑥)(𝑇 − 𝑇𝑚𝑖𝑛)2
(𝑇𝑜𝑝𝑡 − 𝑇𝑚𝑖𝑛)[(𝑇𝑜𝑝𝑡 − 𝑇𝑚𝑖𝑛)(𝑇 − 𝑇𝑜𝑝𝑡) − (𝑇𝑜𝑝𝑡 − 𝑇𝑚𝑎𝑥)(𝑇𝑜𝑝𝑡 + 𝑇𝑚𝑖𝑛 − 2𝑇)]
The parameters are function of the maximum specific growth rate (μm), the rate between light intensity
and optimum light intensity for photosynthesis (i), optimal temperature (Topt), maximum and minimum
temperature where growth is inhibited. These auxiliary functions (μm and Iopt) are all given in the form of
the following correlation, where X is the parameter and γm, αg, To and Topt are constants given by each
type of algae that can be seen in the Table 3 below:
𝑋 (= 𝜇𝑚 𝑎𝑛𝑑 𝐼𝑜𝑝𝑡) = 𝛾𝑚 ∗ (𝑇
𝑇𝑜𝑝𝑡∗ 𝑤𝑢)
𝛼𝑔
(𝑋𝐿)
𝑤 =𝑇 − 𝑇𝑜
𝑇𝑜𝑝𝑡 − 𝑇𝑜 (𝑋𝐿𝐼)
𝑢 =𝑇𝑜 − 𝑇𝑜𝑝𝑡
𝑇𝑜𝑝𝑡 (𝑋𝐿𝐼𝐼)
Table 3: Function parameters for the species evaluated
Oscillatoria agardhii
Phormidium bohneri
μm Iopt μm Iopt
γm 0.5 101 1.7 541
Topt 37 24 37 30
To 43 43 43 43
αg 0.3 1.6 2.7 4.8
-62-
With the evaluation of how the growth rate is reduced, a proportional reduction in maintenance costs of
the current algae growth suppression method will be estimated. This means that in a normal condition
with no water coverage, for example, 100 US$ are spent with prevention measures, in case growth rate is
reduced by 50% this expenditure will be decreased to 50 US$. The worst-case scenario which is the
development of algal blooms will not be evaluated, as it is a situation that is completely out of reach to
predict, as it could happen one or more times in several years or not. The maintenance cost, on the other
hand, is steady and is part of every year’s budget.
6 Results
The results will be divided in two sections. The design section will present results such as energy
generated, size of the power plant, capacity factor and other common utilized parameters to measure the
efficiency of the power plant. Next, the economics section will be analyzed with the input of the design
results to compare if the FPV is more economically feasible than the ground-mounted in a set of
scenarios.
6.1 Design
The overall design parameters are shared by both power plants, meaning that the PV panel and inverter
models, the desired array, DC to AC ratio, orientation, tilt and losses associated to the DC and AC
conversion were kept the same. Table 4 summarizes the design specifications of the PV plants.
Table 4: Common design parameters for FPV and GMPV
Parameter Value
Site Location 38°31'05.0"N 122°48'10.6"W,
California
PV module CanadianSolar CS3U-350P
Inverter Delta M80U_120
DC to AC ratio 1.01
Number of inverters 36
Modules per string 17
Tilt Fixed 33°
Azimuth South-facing (180°)
Average annual soiling loss 5%
Module mismatch 2%
DC wiring losses 2%
AC wiring losses 1%
PV degradation rate 0.5%/year
The first interesting parameter to be assessed is the cell temperature on both power plants. Since both are
under the same weather conditions, cell temperature will be the differential on the energy output. Due to
-63-
the presence of water below the panels, the FPV is expected to operate under lower temperatures, which
is beneficial for the energy output. Figure 42 compares the hourly average cell temperature during
operational hours between the two designs in each of the months.
Figure 42: Hourly average cell temperature for both designs
The plot above shows that cell temperature of the FPV is lower for every month of the year, being
September and December when the highest difference gap occurs, with cell temperatures 17.95% and
17.76% lower, respectively. On the other hand, during the months of June and July, the difference gap
reduces to 12.84% and 12.24%. On a yearly average, the temperature of the floating panels is 14.97%
lower than the ground-mounted counterpart. This lower temperature operation translated to a better
energy output which can be seen distributed for the lifetime of the power plants in Figure 43.
Figure 43: Lifetime electricity generation and capacity factors per year
0
5
10
15
20
25
30
35
40
Tem
per
atu
re (
°C)
Hourly average cell temperature per month
FPV average Ground-mounted average
17.0%
17.5%
18.0%
18.5%
19.0%
19.5%
20.0%
20.5%
21.0%
21.5%
22.0%
4.40
4.60
4.80
5.00
5.20
5.40
5.60
5.80
1 2 3 4 5 6 7 8 9 10111213141516171819202122232425
Cap
acit
y Fa
cto
r
Elec
tric
ity
(GW
h)
Lifetime Electricity Generation
FPV Ground-mounted CF FPV CF Ground-Mounted
-64-
On the first year of operation, the 3 MWp FPV could produce 5.64 GWh compared to the 5.50 GWh of
the GMPV, which corresponds to an increase of 2.54% more electricity on a yearly basis. Considering the
yearly demand of electricity to remain the same, the surplus of around 3.7 GWh could be sold to the grid
on the first year of the operation by the FPV and 3.5 GWh by the GMPV.
This gain in electricity output reflects in a higher capacity factor, which for the FPV reaches 21.47%
compared to 20.94% of the GMPV on the first year of operation. This outcome is in agreement with
previous measurements made in commercial scale FPV in South Korea, with an average increase of 7.6%
[28]. Moving back to the US, between 2014 and 2017, utility ground-mounted solar capacity factors for
the state of California have maintained slightly over 25% [65]. This average is highly influenced by hourly
solar profile, the tracking and concentrating systems, losses and DC to AC ratio, which is usually
oversized for utility solar. In the case of this work’s design, tracking was not utilized and a DC to AC ratio
was set at 1, adding up to the fact that the region is not the sunniest of the state. Utility scale solar usually
tend to be placed on the best location possible, which in the case of California is to the south of the state
near the border with Mexico and neighbouring states Arizona and Nevada. Although holding a capacity
factor lower than large ground-mounted utility solar, both suggested power plants have a higher usage
than commercial (300 kW DC) and residential (5 kW DC) PV, with capacity factors ranging below 20%
for sunnier regions such as Los Angeles [66]. Therefore, the technical modelling has shown reliability with
capacity factors when compared to real values.
Meanwhile, the conversion efficiency increased on average 2.5% for the FPV, which is far from the results
obtained for example in an experiment in Brazil, which increased up to 12.5% [30]. However, most of
efficiency measurements presented in the literature are performed under microscale and experimental
conditions, ranging from cells to modules and, and at different geographies. However, this result attests
that water has a cooling effect on the panels and one could expect an increase in efficiency of the order of
at least 2% when installing floating PV in similar regions in Sonoma county.
To put the above numbers into a practical perspective, the town of Windsor located to the north of the
wastewater facility will serve as an example. The town has around 27,500 inhabitants and some 9,114
households consuming on average 6,536 kWh/year per household [67]. By covering 46.2% of the 70,695
m² water storage pond, the floating solar would be able to provide clean energy for 863 Windsor
households or 9.5% of the customers for a whole year. Moreover, considering the town’s surface area
being 18.89 km², only 0.17% would be used to achieve this energy generation. To provide clean energy to
all the residential customers, 1.82% of the town’s area would have to be covered with PV panels, which
corresponds to five wastewater ponds of equal size as the one chosen for this project. Just Sonoma
County Water Agency alone operates in total three ponds of this size in the city. Since the degradation rate
was considered constant over the years, at the end of the FPV lifetime, it would’ve produced almost 3.5
GWh more than the GMPV or 139.8 MWh per year more. This means that the FPV can power 21 more
households in Windsor per year than the GMPV could.
In terms of the layout, the FPV plant would have a clear advantage. Not only land preparation is not
needed, which includes surface clearing and grubbing of plant roots and land grading, but also building a
foundation is not required. By choosing a pure float, like the one from the Figure 20 (a), enables the
assembly to be done outside of the water and rather faster than of what a ground-mounted would require.
Moreover, the floating design would assume a continuous and single structure, as the area of the pond is
enough for the size of the proposed power plant. In the case of the ground-mounted structure, the
powerplant would have to be divided in two parts, being one half located on the concrete ground next to
the pond proposed for the FPV and a second half located to the southeastern corner of the land. Figure
44 (a) and (b) represents what would be the overview of both power plants.
-65-
(a)
(b)
(c)
Figure 44: (a) Floating PV layout (b) Ground-mounted PV layout (c) High shading on the southeastern part on winter solstice
The disadvantage of the ground-mounted location is visible. The southeastern corner is prone to shading,
as strip of trees exist to the south and buildings to the east. Therefore, given that any other space is
available, one might need to consider what effect these obstacles would have on the energy output of the
set of panels located there. In fact, the shadow cast by the trees on the 23rd of December (winter solstice)
can be seen in Figure 44 (c) which takes up a considerable amount of space that could be used for PVs.
This shows a great advantage for the FPV that could be placed in the middle of the pond and avoid any
shading of surrounding objects. Other alternative would be acquiring a percentage or a whole terrain to
build the GMPV, but this would imply in more costs. This scenario will also be analyzed in the following
sections.
6.2 Economics
The LCOE is one of the parameters that is widely used to compare different energy sources costs and is
normally a measure that tells if a renewable technology has become cost-competitive against the
traditional energy generation. Meanwhile, the NPV can show which project should be chosen given
different investment alternatives. A known trend is that renewables have drawn nearer to fossil fuel
technologies in terms of economic feasibility. Several factors may explain this: reduction in costs coming
from technological improvements and competitive procurement, meaning that new types of market
strategies and financing have enabled the deployment of more projects. In the following subsections, the
different scenarios proposed will be presented and analyzed in detail given its own particularities. Table 5
summarizes common parameters that will be used and will remain fixed along the lifetime analysis of all
the scenarios.
-66-
Table 5: Constant economic parameters
Parameter Value
Discount rate 6.75%
Lifetime of project 25 years
Investment Tax Credit (ITC) 26% (starting year: 2020)
Feed-in-Tariff (FIT) 61.23 US$/MWh
O&M Costs (US$/Wp) 0.013 GMPV/0.026 FPV
Wastewater facility electricity
demand (MWh/year)
1916.5
Price of electricity (US$/kWh) 0.1461
6.2.1 Business-as-Usual scenario
This scenario considers only the usual components of the LCOE and NPV equation, meaning that no
extra benefits related to FPV will be accounted here. For the GMPV, it is considered that the land
available is enough to fit the 3 MWp power plant. The results can be found in Table 6 below.
Table 6: BAU scenario results
FPV GMPV
Without ITC
It (US$) $6,019,143 $4,152,114
LCOE (US$/MWh) 107.88 73.50
NPV (US$) -$1,065,813 $1,168,365
IRR 4.69% 9.74%
Payback (years) N/A 15
With ITC
It (US$) $4,454,166 $3,072,564
LCOE (US$/MWh) 83.58 56.31
NPV (US$) $499,164 $2,247,915
IRR 7.97% 14.12%
Payback (years) 20 10
Here, the results show clearly the importance the federal Investment Tax Credit has on the viability of the
FPV. Without this incentive, the project would be considered unfeasible, because it would not return the
investment made during the lifetime considered. On the other hand, the GMPV results show that costs
have come down in such a way that the project would still be feasible for this location without such
incentives. When the federal subsidy is considered, the floating PV becomes just enough to be feasible, as
it can be seen by the IRR being close to the discount rate considered (6.75%). Meanwhile, the GMPV
project shows the highest profitability and the lowest payback time. The LCOE results retrieved for the
GMPV are in accordance with benchmarks performed in regions such as Dagget in California and
Arizona, which have more favourable conditions than Windsor. For these regions, baseline commercial
plants between 200-300 kW have LCOEs ranging from 60 US$/MWh to 90 US$/MWh with and without
ITC, respectively. For large utility solar over 100 MW, the LCOE reaches between 40 to 30 US$/MWh
[55], [66]. Since the proposed project is 3 MWp, it sits in between these benchmark ranges which
reinforces the realistic results of the model. For the FPV, LCOE hasn’t still been calculated on a world
average or even for the North-American reality, due to the limited number of projects deployed. The most
-67-
recent documented LCOE calculations refer to the coupling of FPVs with hydropower dams in Brazil
which present values depending on the tilt of the panels. For the a tilt equal to the one proposed in this
project (33°), the authors found LCOEs ranging from 119 to 132 US$/MWh [68]. Nonetheless, these
values do not tell much about how precise the thesis’ LCOE results are, but rather how accurate. FPVs
LCOE are between 40 to 50% higher than a ground-mounted PV in Brazil, which is also in accordance to
the results of this work.
These results are expected since the floating PV haven’t had enough tracking history and the learning
curve does not quite follow the regular PV projects. Although these two projects are both based on
photovoltaic energy, the mounting and deployment is completely different, being the FPV closely related
to marine installations because of the anchoring and mooring technology. As it was seen before, this
component of the structure may be responsible for 20% of the overall installation costs, depending on the
environment conditions it must withstand. Understanding the costs of anchoring is a challenge itself, since
it depends on the type of soil, type of anchoring, weather conditions (wind, wind direction, waves, water
level variation). In this project, the floating structure was considered to be anchored at the bottom and the
impact on the costs makes up close to 20% of the total investment. If a different type of anchoring is
chosen, one simplifying the construction phase, and together with a decrease in costs of the floating
structures, the NPV of the FPV will approach the NPV of the GMPV. Figure 45 shows the sensitivity
analysis performed for the FPV, decreasing the installations costs until the NPV becomes positive.
Figure 45: Sensitivity analysis of NPV varying the structural costs of the FPV
Maintaining all factors equal, but decreasing the structural costs mainly related to the floats and
anchoring/mooring, the NPV of the FPV will approach the profitability of a GMPV. The main difference
will stay on the O&M costs which are still an incognita for floating solar.
As an extended simple analysis, if tracking were to be considered for the GMPV, one could expect the
same land-use intensity and an overall increase in energy yield of around 18% (recall Figure 8 for San
Francisco, CA). Therefore, having the same land impact as the fixed tilt GMPV, but installation costs as
high as 20% more of the fixed tilt alternative [10], the GMPV with tracking would still fare better than the
floating PV in this scenario but would not make a huge impact in terms of LCOE and NPV (58.8
US$/MWh and 2.36 MUS$).
-$2
-$1
-$1
$0
$1
$1
$2
$2
$3
0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80
NP
V (
mill
ion
s U
S$)
Structural Costs (US$/Wp)
NPV sensitivity analysis with structural costs
FPV no ITC FPV with ITC GMPV no ITC GMPV with ITC
-68-
6.2.2 Extra benefits scenario
The extra benefits scenario will try to merge both proven advantages of using solar power with the
supposed benefits that covering a water pond with PV panels might have. In this case, it will account for
the gains in evaporation translated to extra sales of water and a decrease in maintenance costs related to
algae.
The proposed floating power plant will only cover 46.18% of the pond’s surface area due to the PV
capacity restriction being an eligibility criterion for the ReMAT program in California. The modelling
results show that for the region chosen, the rate of evapotranspiration is 71.21 cm/year. The profile of
daily evaporation can be seen in Figure 46.
Figure 46: Daily evaporation rates of the chosen pond
The calculated rate results in the evaporation of 50,341 m³ of water per year for the pond. These results
are in agreement with previous studies showing an estimated 61 to 70 cm/year of evapotranspiration for
Sonoma County, as it can be observed in Figure 47.
Figure 47: Estimated mean annual evapotranspiration for the US, Sonoma County highlighted by a red circle. Adapted from [69]
0
1
2
3
4
5
6
7
8
9
0 50 100 150 200 250 300 350
Evap
ora
tio
n (
mm
/day
)
Day of the year
Evaporation profile of the pond
-69-
It was assumed that covering a specific percentage of surface with floating solar could potentially reduce
the evaporation of this surface to zero. Hence, the Agency could save up to 23,250 m³ of water per year
just by placing the floating structure over water. As previously mentioned, the treated water of the chosen
pond is mostly used for the irrigation of wineries. In Sonoma County alone, 58,280 acres are vineyards, in
which 40% of them are of 20 acres or less. Considering a 20-acre winery consuming 12 acre-foot per year
(14,800 m³/year) [70], the water saved by placing the FPV over the pond would correspond to supplying a
winery of this size for one and half year. Therefore, considering that a winery of this dimensions is setup
in the next year, Sonoma County Agency would secure an extra of at least 18,867 US$ in revenues in the
first year of operation.
As for algal control, the modelling of this benefit was largely simplified, since many types of algae exist
and the models have yet to be proven on large scale. For this work, two types of algae commonly found
on freshwater ponds were chosen and their behaviour simulated. First, it was considered that for the strip
of area where the FPV is placed, 95% of light intensity is reduced. However, the FPV only covers 46% of
the pond, which means that the other 54% of the pond receive the normal light intensity coming from the
atmosphere. With the problem framed in this way, one can assess how the partial cover of the floating PV
will impact the algal growth as a whole. Figure 48 shows the behaviour of Oscillatoria agardhii and
Phormidium bohneri with and without partial cover of the FPV.
Figure 48: Algal growth rates with no cover and partial FPV cover
Since light has been shaded for part of the pond, algae can still grow, but the growth rate will occur at
lower light intensities. For the case of these two species, yearly average growth rate increases 9.95% for
Oscillatoria agardhii and 0.36% for Phormidium bohneri. This differs from the advertised benefit of algae
growth reduction. What seems to be forgotten when addressing this benefit is that not only light is a
determinant factor in the photosynthesis of algae. In fact, temperature plays a major role in the growth
rate and, if kept as it is, may influence negatively. Figure 49 exemplifies the behavior of two types of algae
that, even with less light, still have a higher growth rate at certain temperature ranges. Figure 50 shows the
sensitivity analysis of the model results of Oscillatoria agardhii framed in the same type of plot.
0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.00 500.00 1000.00 1500.00 2000.00
Gro
wth
rat
e((1
/day
)
Light Intensity (μmol/m²/s)
Oscillatoria agardhii growth rate
No cover Partial FPV cover
0.00
0.10
0.20
0.30
0.40
0.50
0.60
0.70
0.80
0.90
0.00 500.00 1000.00 1500.00 2000.00
Gro
wth
rat
e((1
/day
)
Light Intensity (μmol/m²/s)
Phormidium bohneri growth rate
No cover Partial FPV cover
-70-
Figure 49: Growth rate of N. oceanica with different temperature and light intensities [62]
Here, it can be observed that between 10 up to 25 °C, the lowest light intensity gives the highest growth
rate. However, the difference in growth rate is less than 0.1 day-1 which is very small
Figure 50: Oscillatoria agardhii growth rate in fuction of temperature
In the case of Oscillatoria agardhii, the difference in light intensity can result in a growth rate difference of
less than 0.05 day-1 with an increase in temperature. Therefore, given that the changes in growth rates for
the types of algae analyzed are neutral to negative, it will be considered that this benefit is not relevant for
the NPV of the project as costs of maintenance would not decrease.
Therefore, adding up only the extra revenue per year due to evaporation gains, the results can be seen in
Table 7. Here, one hypothesis was done which was the adjustment of the water rates by 4% on yearly
basis, which is set below the 4.75% average from the 20-year evolution of prices (recall water prices on
Figure 41), excluding the years of 2002 and 2010 that increased the rates 16.41% and 27.95%, respectively.
Therefore, the starting Sonoma Aqueduct price of 0.8116 US$/m³ on year 1 would become 2.080 US$/m³
on year 25. The results of the GMPV remain the same and are added in this table only to make the
comparison easier to observe.
0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0.40
0.45
0 5 10 15 20 25 30 35 40 45
Gro
wth
rat
e((1
/day
)
Temperature (°C)
Oscillatoria agardhii growth rate
50 100 200 300 400 500 600 700 800 900
-71-
Table 7: Extra benefit scenario results
FPV GMPV
Without ITC
NPV (US$) -$737,005 $1,168,365
IRR 5.36% 9.74%
Payback (years) N/A 15
With ITC
NPV (US$) $827,972 $2,247,915
IRR 8.73% 14.12%
Payback (years) 18 10
In this scenario, again is possible to observe that the ITC is important for the feasibility of the FPV. When
this subsidy is considered, the project becomes positive with an IRR of 8.73% which is already
considerable higher than the discount rate.
On one hand, compared to the GMPV, the floating power plant does not show the same level of
profitability even though additional benefits have been added. On the other hand, this scenario proves
that adding up the benefit of extra revenues coming from water sales may not only improve the IRR, but
also the payback, reducing it by two years. The net cashflow also improved by an average of 7.73% per
year over the lifetime of the investment, peaking at 15.56% on the year 25 when compared with the
previous FPV scenario. If a company decides to analyze a project under 20 years of lifetime, including the
benefit of the water evaporation can be decisive.
6.2.3 GMPV with land acquisition scenario
This last scenario points out a situation that is a common problem for solar plants where land costs are
high. In countries like Japan, for example, land acquisition corresponds to a large factor on the upfront
costs of a PV plant, which explains the high penetration of the FPV technology there. In the state of
California, land costs have been increasing in the past few years and technological hubs such as San
Francisco and other big centers such as Sacramento have shown above average increases. Figure 51 (a)
shows the average price of an acre of land in the state of California in 2017.
-72-
(a)
(b)
Figure 51: (a) Average land price per acre in California by county in 2017 [71] (b) Benchmarked lands [72]
In the figure above, Sonoma County prices have reached more than a million dollars per acre in 2017.
However, since this is an average, it is highly influenced by the high prices of agricultural land, and
specially, the land that is proper to growing grapes for the winery industry. Thus, since the area needed for
the PV plant is roughly 8 acres, using these average county costs would make a project of this size
unfeasible for any ground-mounted installation. The objective here should be finding a land that is not far
away from the waste water facility, as most of the renewable energy projects from the Sonoma County
Agency are located inside their properties, and acquiring it by a price way below the average of Figure 51
(a). Therefore, a benchmark research on 55 lots of land was done around the cities of Windsor and Santa
Rosa which surround the facility. The areas researched are shown in Figure 51 (b) and have a market
average cost of 225,400 US$/acre or 55.7 US$/m².
Although this PV project is eligible for the solar ITC, this tax credit does not cover land acquisition costs
which will depend on the amount of land that is required. Therefore, a sensitivity analysis was performed
with the acquisition of areas ranging from 5% to 100%. This means that for the first case, 5% of the
power plant would be installed in an external acquired land and the remaining 95% would be placed inside
the boundaries of the existing area. When this becomes 100%, it means that the whole GMPV would be
placed in lands outside of the boundaries of this problem. Since the land to be acquired is probably not
ready for the installation of a large PV project, it might need site preparation works such as clearing and
grubbing, soil stripping, stockpiling, grading and other civil works. For site preparation costs, it will be
considered 25,000 US$/acre [73], with these activities being eligible for the ITC subsidy. The results of the
area sensitivity analysis on the LCOE can be observed in Figure 52.
-73-
Figure 52: LCOE evolution with the increase in external land required compared to FPV’s LCOE
In the figure above, it is possible to observe that if the ITC subsidy is not taken into account, the LCOE
of the GMPV will always be lower than the FPV. On the other hand, if the tax credit is included, the
LCOE of the FPV will become competitive when 85% of extra land is required to place the GMPV.
These values will show the competitiveness in cost for both of technologies, but it won’t mean that the
projects along this chart will be profitable. For this analysis, a sensitivity was also done for the NPV of the
GMPV with the percentage of extra land required, that can be seen in Figure 53.
Figure 53: NPV evolution with the increase in external land required compared to FPV’s NPV
From the figure above, several important observations can be done:
• An unsubsidized ground-mounted solar plant would stay feasible in this location until it would be
necessary acquiring 55% of external land (17,956.44 m²) to be able reach the proposed capacity.
Nonetheless, the GMPV is always more profitable than a FPV in this case;
0
20
40
60
80
100
120
0% 20% 40% 60% 80% 100% 120%
LCO
E (U
S$/M
Wh
)
Percentage of extra land required
Ground-mounted PV LCOE evolution with land acquisition
FPV no ITC FPV with ITC GMPV no ITC GMPV with ITC
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
2.0
2.5
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
NP
V (
mill
ion
US$
)
Percentage of extra land required
Ground-mounted PV NPV evolution with land acquisition
FPV extra benefits no ITC FPV extra benefits with ITC GMPV no ITC
GMPV with ITC FPV no benefits with ITC
GMPV more profitable
FPV more profitable
GMPV more competitive
FPV
more
competitive
-74-
• When the subsidy is accounted, a ground-mounted PV will always be profitable, independently of
the external area required to build. If 100% of the area (32,648.07 m²) is purchased, the NPV will
still remain positive;
• When ITC is included, FPV considering extra benefits will become more profitable than a GMPV
when 75% of external area is required to mount the GMPV. When 70% of external area is
required, the NPV of both projects are very close, with a slight advantage for the ground-
mounted PV. In fact, if a profitability index is calculated, which is the ratio between NPV and
upfront costs, the GMPV with 70% external area has a ratio of 0.1869 and the FPV has 0.1859,
which, although larger, is not a relevant difference. Therefore, it may be concluded that even
when 70% of external area is needed, the FPV might be considered more profitable;
• When ITC is included, FPV not considering any extra benefits will become more profitable than
a GMPV when 90% of external area is required. When 85% of external area is required, the NPV
of both projects are very close, with a slight advantage for the ground-mounted PV. Again, if a
profitability index is calculated, the GMPV with 85% external area has 0.1131 and the FPV has
0.1121. Similar as above, the difference is not relevant, thus it may be concluded that even when
85% of external area is needed, the FPV with no extra benefits might be more profitable;
• FPV without ITC and not considering extra benefits was not included in the chart, as it holds the
most negative NPV.
6.2.4 Summary of results
Three different scenarios were analyzed in the previous sections to understand how floating PV have
evolved in costs. The sum up chart for LCOE for all the scenarios can be observed in Figure 54. The
most competitive technologies should be positioned to the left, whereas the least competitive is located to
the right.
Figure 54: LCOE results for the analyzed scenarios
Since LCOE does not account for the expected revenues of the power plant, LCOE from floating PV
with and without extra benefits are the same. For the chosen location, without a financial incentive, FPV
would be less competitive against GMPV regardless of the amount of land purchased. Although it has
been observed a 2.54 % increase in electricity output on a yearly basis, this does not overcome the upfront
costs of the FPV, that are still more than double than the GMPV.
When ITC is considered, FPV enters the benchmarked range zone for GMPV commercial projects up to
1 MW, which accounts for all the states of the US. However, on a general basis GMPV is usually more
-75-
competitive, even when 85% of external area is required to build. Finally, GMPV is competitively
unbeatable when external land is not required and the federal incentive is considered. Therefore, the
ground-mounted layout can be concluded as the best if, for the location chosen, the area available could
be used for building it. This conclusion also translates to the NPV summary in Figure 55, where profitable
investments are located to the right of the vertical dotted line, whereas non-profitable to the left.
Figure 55: NPV results for the analyzed scenarios
On the analysis of the NPV, the revenues become important and the extra benefits considered make a
slight difference. Nonetheless, federal incentives still will play an important part in the profitability of solar
installations, being it floating or ground-mounted.
FPV are profitable when the federal incentive is considered; furthermore, the scenario that considers the
evaporation gains have an increase in NPV and a decrease in payback of 2 years, which can be decisive
depending on the lifetime that is considered for the investment to be acceptable. However, the floating
option does not beat the GMPV without land acquisition, showing that the track record and reduced costs
of this technology still overpower the extra revenue and efficiency from FPV. Although ITC does not
cover area purchase, GMPV eligible for this incentive have a positive NPV even when the whole project
has to be built in a different location. The area issue needs to be analyzed with caution for other projects,
as the procurement of external land will vary and may impact the results differently. For this thesis, the
nearby lands considered were located in the Central Valley of California, which is a favorable location for
agriculture development in the state. Considering land purchase and federal incentives, GMPV is still
more profitable than FPV up to the purchase of 70% of external land when extra benefits are considered
and 85% when they are not.
7 Conclusions
Solar energy has come along years of development and cost reductions that have made it to be
competitive in today’s market. Besides being a free resource, the generation of electricity with solar does
not pollute and is in line with the sustainable development proposed by several international summits and
agreements. In this work, several layouts have been presented and a techno-economic analysis of floating
and ground-mounted PV was performed for a wastewater treatment facility in California, USA. A
-76-
technical model was developed to estimate the energy yield of a floating PV power plant and, with these
results, three different scenarios were proposed and analyzed to understand the economic feasibility of
this technology compared to the traditional ground-mounted PV.
Regarding the technical aspects, it was found that FPV is in fact more efficient than GMPV of the same
size, which in this work was 3 MWp. This could be observed due to the overall increase of 2.54% more
energy produced per year by the FPV, which in turn would provide an accumulated surplus of 3.5 GWh in
25 years. Therefore, the capacity factor observed for the FPV was 21.47% in the first year of operation,
which is slightly superior than benchmarked ground-mounted systems in California that range below 20%.
This performance was the result of the 15% yearly average decrease in cell temperature of the floating
panels. On a qualitative analysis, the FPV layout for the location chosen shows several advantages. Land
preparation is not required, which can add up to 25,000 US$/acre in costs, as well as no need for
foundation and other civil works. Moreover, the construction and deployment of FPV are made easier and
faster if modular floats are used, which can be assembled outside of the pond and be pushed inside to the
desired position. Finally, FPV can be placed all together, whereas GMPV has two alternatives: dividing the
power plant in two separate pieces of land or purchasing extra land.
In terms of economic feasibility, three scenarios were proposed and analyzed. The first scenario (Business-
as-Usual) considered the FPV and GMPV only accounting for total upfront costs (excluding land
purchasing), maintenance, energy and revenues to calculate the LCOE and NPV. Here, it was found that
even without federal incentives, the ground-mounted design would be profitable and more competitive
than the FPV in terms of LCOE. When the Investment Tax Credit (ITC) is included, floating solar
becomes profitable, although much less than the ground-mounted project.
The second scenario included the usually advertised extra benefits of the floating PV, which are
evaporation savings and algae growth suppression. For a wastewater facility, both benefits could easily
lead to economic gains, as evaporation of treated water means lost revenue and the growth of unwanted
algae mean an increase in treatment expenditure. However, it was found that algae growth wasn’t really
affected by covering part of the pond with a floating structure. In fact, for some types of algae, the growth
rate might increase when less light is cast, given an optimal temperature. In view of this, only the water
that would not evaporate from the covered part of the pond, was considered an extra revenue. By opting
to place the 3 MWp FPV over the pond, Sonoma County Water Agency would save up to 23,250 m³ of
water per year which represents, on the first year of operation, 18,867 US$ on extra revenues. Hence,
when comparing FPV without extra benefits and with benefits, the NPV will almost double, however,
without the federal incentive the NPV would still lie below zero. Although the NPV improves, the
ground-mounted design would still be more interesting for the location.
The last scenario accounted for the need to purchase extra land to build the whole or part of the GMPV.
A benchmark on the negotiated land prices around the cities of Windsor and Santa Rosa showed an
average price of 225,400 US$/acre or 55.7 US$/m². Considering the federal incentive, the GMPV proved
to be profitable even when 100% of the area needs to be purchased, but when 70% of external land is
required, the FPV with extra benefits shows to be a more interesting project. When the extra benefits of
the FPV are not considered, this change is observed when the 85% of external land is purchased.
Ultimately, the FPV has proven to be more efficient and practical for the location chosen. Overall, the
LCOE of the proposed FPV is competitive with the GMPV considering land purchase and with the
benchmarked ground-mounted projects around the USA. Payback times ranged from 18 to 20 years in
FPV considering extra benefits and without extra benefits, respectively. If the lifetime of a project is
considered 20 years rather than 25, the evaporation gains could be relevant in a feasibility analysis.
However, for the location chosen, if the available area isn’t going to be utilized for any other attribution
and, thus, no extra land is required to be purchased, the ground-mounted design will be the better option
both in LCOE and NPV.
-77-
8 Limitations and Suggestions for future work
Based on the work presented, several limitations were identified throughout the analysis and could be
further improved in future work. Given the limited literature availability and track record of floating PV
projects, most of the suggestions will lie on technical aspects. Improvements in future analysis should
account for:
• Modelling of a cell temperature equation accounting for real data, mainly measuring the water
temperature and energy yield of a floating installation;
• Wind drag and lift forces on the panels were not considered in this work. Since the panels were
set at 33° tilt, one must account for the forces that wind would exert on the panels. This would
also influence the anchoring system, that would be under more force, and potentially increasing
the costs;
• Most commercial FPV are placed on a 12° tilt where the shade and interrow spacing is
minimized. This means that in between rows, only one float needs be used to create a walking
path, to enable maintenance works. When the panel is tilted to a higher angle, the interrow space
is bigger to avoid unwanted shading. Therefore, the walking path between rows could require
extra floats. Since this work was aimed at optimizing the kWh/kWp, it hasn’t accounted for the
extra floats that might be required and how this could affect upfront costs;
• Soiling losses are believed to be smaller in FPV than in GMPV. In this work the losses were the
same, due to insufficient proof of this belief. Therefore, an analysis on how ground and water
environment may affect soiling is required;
• The same reasoning is done for the degradation of the panels. It was considered that both have
the same constant degradation through the years, due to the lack of evidence on the FPV.
-78-
References
[1] L. M. Ross, “How the 1973 Oil Embargo Saved the Planet,” Foreign Affairs, 2013. [Online]. Available: https://www.foreignaffairs.com/articles/north-america/2013-10-15/how-1973-oil-embargo-saved-planet. [Accessed: 10-Feb-2019].
[2] EIA, “International Energy Outlook 2017,” 2017.
[3] M. Fagan, “Sheikh Yamani predicts price crash as age of oil ends,” 2000. [Online]. Available: https://www.telegraph.co.uk/news/uknews/1344832/Sheikh-Yamani-predicts-price-crash-as-age-of-oil-ends.html. [Accessed: 28-Dec-2018].
[4] T. Embury-Dennis, “Costa Rica’s electricity generated by renewable energy for 300 days in 2017,” 2017. [Online]. Available: https://www.independent.co.uk/news/world/americas/costa-rica-electricity-renewable-energy-300-days-2017-record-wind-hydro-solar-water-a8069111.html. [Accessed: 12-Dec-2018].
[5] R. Perez and M. Perez, “A Fundamental Look At Supply Side Energy Reserves For The Planet,” IEA-SHCP-Newsletter, vol. 62, no. 1, pp. 2–5, 2015.
[6] G. M. Masters, Renewable and Efficient Electric Power Systems. Hoboken, New Jersey: JOHN WILEY & SONS, INC., 2004.
[7] A. Sahu, N. Yadav, and K. Sudhakar, “Floating photovoltaic power plant: A review,” Renew. Sustain. Energy Rev., vol. 66, pp. 815–824, 2016.
[8] SEIA, “Solar Means Business Tracking Corporate Solar Adoption in the U.S.,” 2018.
[9] B. Hoen et al., “Selling Into the Sun: Price Premium Analysis of a Multi-State Dataset of Solar Homes,” 2015.
[10] S. Ong, C. Campbell, P. Denholm, R. Margolis, and G. Heath, “Land-Use Requirements for Solar Power Plants in the United States,” 2013.
[11] P. Heinstein, C. Ballif, and L. E. Perret-Aebi, “Building integrated photovoltaics (BIPV): Review, potentials, barriers and myths,” Green, vol. 3, no. 2, pp. 125–156, 2013.
[12] O. Shukla, B. B. Jani, S. M. Agravat, N. Srivastava, and G. Singh, “Canal Top Solar PV Plant in Gujarat,” 2016.
[13] Swimsol, “96kWp – SolarSea®, Baa Atoll, Maldives,” 2016. [Online]. Available: https://swimsol.com/solar-projects/floating-photovoltaic-offshore-solar-sea-power-pv-modular-96kwp/. [Accessed: 26-Dec-2018].
[14] A. Deutsch, “Dutch plan to build giant offshore solar power farm,” 2018. [Online]. Available: https://www.reuters.com/article/netherlands-solar-offshore/dutch-plan-to-build-giant-offshore-solar-power-farm-idUSL8N1Q46M0. [Accessed: 28-Dec-2018].
[15] Autostrada del Brennero SpA, “NOISE PROTECTION WALL With integrated solar panels.,” 2016. [Online]. Available: https://www.autobrennero.it/en/the-motorway-network/sustainable-development/noise-protection-wall-/. [Accessed: 26-Dec-2018].
[16] World Bank Group, ESMAP, and SERIS, “Where sun meets water floating solar market report,” 2018.
[17] M. Smyth, J. Russell, and T. Milanowski, Solar Energy in the Winemaking Industry, 1st ed. Springer, 2011.
[18] IHA, “Case study: A hybrid hydropower and floating PV system in Portugal,” London, 2018.
[19] J. Weaver, “Floating solar power gets 10 year fixed lease for California municipality,” PV Maganize, 2018. [Online]. Available: https://pv-magazine-usa.com/2018/09/19/floating-solar-has-a-nice-summer-going-on-252-kw-completed-5-mw-coming-for-ciel-terre/. [Accessed: 07-Jan-2019].
-79-
[20] CBS, “Walden Harnesses The Power Of The Sun With Floating Solar Panels,” 2018. [Online]. Available: https://denver.cbslocal.com/2018/10/19/walden-solar-power-floating-panels/. [Accessed: 30-Dec-2018].
[21] R. S. Spencer, J. Macknick, A. Aznar, A. Warren, and M. O. Reese, “Floating PV: Assessing the Technical Potential of Photovoltaic Systems on Man-Made Water Bodies in the Continental U.S.,” Environ. Sci. Technol., p. acs.est.8b04735, 2018.
[22] EPA, “Principles of Design and Operations of Wastewater Treatment Pond Systems for Plant Operators, Engineers, and Managers,” Cincinnati, 2011.
[23] Legislative Council of the Hong Kong Special Administrative Region, “Feed-in tariff for solar power in selected places,” 2017.
[24] T. Kenning, “Taiwan ripe for floating PV,” PVTech, 2016. [Online]. Available: https://www.pv-tech.org/news/taiwan-ripe-for-floating-pv. [Accessed: 04-Jan-2019].
[25] Energysage, “Are solar panels resistant to salt and corrosion?,” 2017. [Online]. Available: https://news.energysage.com/solar-panels-resistant-salt-corrosion/. [Accessed: 30-Dec-2018].
[26] EVIVA, “Come costruire un pannello solare termico fai-da-te,” 2018. [Online]. Available: https://www.evivaenergia.com/news/2018/02/come-costruire-un-pannello-solare-termico-fai-da-te/. [Accessed: 30-Dec-2018].
[27] Sinovoltaics, “Floating Solar (PV) Systems: why they are taking off,” 2015. [Online]. Available: https://sinovoltaics.com/technology/floating-solar-pv-systems-why-they-are-taking-off/. [Accessed: 30-Dec-2018].
[28] Y. K. Choi, “A study on power generation analysis of floating PV system considering environmental impact,” Int. J. Softw. Eng. its Appl., vol. 8, no. 1, pp. 75–84, 2014.
[29] W. C. L. Kamuyu, J. R. Lim, C. S. Won, and H. K. Ahn, “Prediction model of photovoltaic module temperature for power performance of floating PVs,” Energies, vol. 11, no. 2, 2018.
[30] E. M. do Sacramento, P. C. M. Carvalho, J. C. de Araújo, D. B. Riffel, R. M. da C. Corrêa, and J. S. Pinheiro Neto, “Scenarios for use of floating photovoltaic plants in Brazilian reservoirs,” IET Renew. Power Gener., vol. 9, no. 8, pp. 1019–1024, 2015.
[31] H. Liu, V. Krishna, J. Lun Leung, T. Reindl, and L. Zhao, “Field experience and performance analysis of floating PV technologies in the tropics,” Prog. Photovoltaics Res. Appl., vol. 26, no. 12, pp. 957–967, 2018.
[32] M. Rosa-Clot and M. G. Tina, SUBMERGED AND FLOATING PHOTOVOLTAIC SYSTEMS Modelling, Design and Case Studies. Academic Press, 2017.
[33] Ciel et Terre, “Ciel et Terre,” 2018. [Online]. Available: https://www.ciel-et-terre.net. [Accessed: 29-Dec-2018].
[34] LSIS, “Floating Photovoltaic System,” 2016. [Online]. Available: http://www.lsis.com/edm/2016/1128/1611_electric_Floating Photovoltaic System.pdf. [Accessed: 28-Dec-2018].
[35] LG, “Floating PV Design & Quality Challenges,” 2017. [Online]. Available: https://www.ibesalliance.org/fileadmin/content/download/presentation/2017/JF4S_Intersolar_Middle_East/2._Veyis_Neo_TOPRAK.pdf. [Accessed: 28-Dec-2018].
[36] Solaris Synergy, “Solaris Synergy (Israel) Ltd. A leader in the provision and operation of Floating PV Solar system s,” no. June, 2015.
[37] N. Abdel-Raouf, A. A. Al-Homaidan, and I. B. M. Ibraheem, “Microalgae and wastewater treatment,” Saudi J. Biol. Sci., vol. 19, no. 3, pp. 257–275, 2012.
[38] P. Beiter, M. Elchinger, and T. Tian, “2016 Renewable Energy Data Book,” 2016.
[39] DSIRE, “Database of State Incentives for Renewables & Efficiency®,” 2018. [Online]. Available: http://www.dsireusa.org. [Accessed: 06-Jan-2019].
-80-
[40] R. Margolis, D. Feldman, and D. Boff, “Solar Industry Update,” Natl. Renew. Energy Lab., no. May, pp. 1–83, 2018.
[41] A. Grennell, “Why 96 million plastic ‘shade balls’ dumped into the LA Reservoir may not save water,” PBS, 2018. [Online]. Available: https://www.pbs.org/newshour/science/why-96-million-plastic-shade-balls-dumped-into-the-la-reservoir-may-not-save-water. [Accessed: 07-Jan-2019].
[42] M. S. Kannan and N. Lenca, “Field guide to algae and other ‘scums’ in ponds , lakes , streams and rivers,” 2013.
[43] M. Wright, “Problems In Wainscott Pond Are Diverse, But Solutions May Be Quick,” The Southampton Press, 2019. [Online]. Available: http://www.27east.com/news/article.cfm/Wainscott/581040/Problems-In-Wainscott-Pond-Are-Diverse-But-Solutions-May-Be-Quick. [Accessed: 26-Jan-2019].
[44] H. Wang et al., “Comparative analysis of energy intensity and carbon emissions in wastewater treatment in USA, Germany, China and South Africa,” Appl. Energy, vol. 184, pp. 873–881, 2016.
[45] EIA, “Table 5.6.A. Average Price of Electricity to Ultimate Customers by End-Use Sector, by State,” 2018. [Online]. Available: https://www.eia.gov/electricity/monthly/epm_table_grapher.php?t=epmt_5_6_a. [Accessed: 08-Jan-2019].
[46] Sonoma County Water Agency, “Airport-Larkfield-Wikiup Sanitation Zone Sewer System Management Plan (SSMP),” 2016.
[47] NREL, “System Advisor Model (SAM),” 2018. [Online]. Available: https://sam.nrel.gov. [Accessed: 08-Jan-2019].
[48] United States Bureau of Reclamation, “Floatovoltaics® Solar Power System Overview and SPG Solar Statement of Qualifications,” 2011. [Online]. Available: https://www.usbr.gov/lc/region/programs/crbstudy/1_Evaporation_Reduction_via_Floatovoltaics_Systems.pdf. [Accessed: 09-Jan-2019].
[49] Sonoma County Water Agency, “FY 2018-19 BUDGET BUDGET SECTION SUMMARY,” 2018. [Online]. Available: https://www.sonomawater.org/ALWSZ/. [Accessed: 09-Jan-2019].
[50] K. Zipp, “Why array oversizing makes financial sense,” 2018. [Online]. Available: https://new.abb.com/docs/librariesprovider117/default-document-library/solar-inverters/solar_power_world-article.pdf?sfvrsn=80a7614_4. [Accessed: 18-Jan-2019].
[51] P. Gilman, “SAM Photovoltaic Model Technical Reference Update,” Natl. Renew. Energy Lab. NREL P-6A20-64102, vol. 3, no. 1, pp. 1–93, 2018.
[52] O. Mohseni, H. G. Stefan, and T. R. Erickson, “A nonlinear regression model for weekly stream temperatures,” Water Resour. Res., vol. 34, no. 10, pp. 2685–2692, 1998.
[53] MathWorks, “MATLAB,” 2019. [Online]. Available: https://www.mathworks.com/products/matlab.html. [Accessed: 19-Jan-2019].
[54] M. Hejri, H. Mokhtari, M. R. Azizian, and L. Söder, “An analytical-numerical approach for parameter determination of a five-parameter single-diode model of photovoltaic cells and modules,” Int. J. Sustain. Energy, vol. 35, no. 4, pp. 396–410, 2016.
[55] R. Fu, D. Feldman, R. Margolis, R. Fu, D. Feldman, and R. Margolis, “U . S . Solar Photovoltaic
System Cost Benchmark : Q1 2018 U . S . Solar Photovoltaic System Cost Benchmark : Q1 2018,” no. November, 2018.
[56] M. Redón Santafé, J. B. Torregrosa Soler, F. J. Sánchez Romero, P. S. Ferrer Gisbert, J. J. Ferrán Gozálvez, and C. M. Ferrer Gisbert, “Theoretical and experimental analysis of a floating photovoltaic cover for water irrigation reservoirs,” Energy, vol. 67, pp. 246–255, 2014.
[57] Grant Thornton, “Renewable energy discount rate survey results – 2017,” 2018. [Online]. Available: http://www.cleanenergypipeline.com/Resources/CE/ResearchReports/renewable-energy-discount-rate-survey-2017.pdf. [Accessed: 02-Feb-2019].
-81-
[58] L. Zotarelli, M. D. Dukes, C. C. Romero, K. W. Migliaccio, and T. Kelly, “AE459/AE459: Step by Step Calculation of the Penman-Monteith Evapotranspiration (FAO-56 Method),” pp. 1–10, 2013.
[59] N. An, S. Hemmati, and Y. J. Cui, “Assessment of the methods for determining net radiation at different time-scales of meteorological variables,” J. Rock Mech. Geotech. Eng., vol. 9, no. 2, pp. 239–246, 2017.
[60] Sonoma County Water Agency, “Water Rates for Fiscal Year 2018-2019.” Santa Rosa, p. 79, 2018.
[61] P. Talbot, J. M. Thébault, A. Dauta, and J. de la Noüe, “A comparative study and mathematical modeling of temperature, light and growth of three microalgae potentially useful for wastewater treatment,” Water Res., vol. 25, no. 4, pp. 465–472, 1991.
[62] O. Bernard and B. Rémond, “Bioresource Technology Validation of a simple model accounting for light and temperature effect on microalgal growth,” Bioresour. Technol., vol. 123, pp. 520–527, 2012.
[63] Berthold Technologies, “How do I convert irradiance into photon flux?” [Online]. Available: https://www.berthold-bio.com/service-support/support-portal/knowledge-base/how-do-i-convert-irradiance-into-photon-flux.html. [Accessed: 27-Jan-2019].
[64] Algae Research and Supply, “LIGHTING FOR ALGAE CULTURES.” [Online]. Available: https://algaeresearchsupply.com/pages/lighting-for-algae-cultures. [Accessed: 27-Jan-2019].
[65] EIA, “Electric Power Monthly,” 2019. [Online]. Available: https://www.eia.gov/electricity/monthly/epm_table_grapher.php?t=epmt_6_07_b. [Accessed: 16-Feb-2019].
[66] NREL, “Annual Technology Baseline,” 2018. [Online]. Available: https://atb.nrel.gov/electricity/2018/index.html?t=sd. [Accessed: 16-Feb-2019].
[67] Pacific Gas and Electric, “PG&E Energy Data Request-Public Data Sets,” 2018. [Online]. Available: https://pge-energydatarequest.com/public_datasets/download?type=electric&file=PGE_2018_Q4_ElectricUsageByZip.zip. [Accessed: 13-Feb-2019].
[68] N. Motta et al., “Use of floating PV plants for coordinated operation with hydropower plants : Case study of the hydroelectric plants of the São Francisco River basin,” Energy Convers. Manag., vol. 171, no. May, pp. 339–349, 2018.
[69] W. E. Sanford and D. L. Selnick, “ESTIMATION OF EVAPOTRANSPIRATION ACROSS THE CONTERMINOUS UNITED STATES USING A REGRESSION WITH CLIMATE AND LAND-COVER DATA,” J. Am. Water Resour. Assoc., vol. 49, no. 1, pp. 217–230, 2013.
[70] Sonoma County Winegrowers, “GRAPE GROWING & VINEYARD FACT SHEET,” 2015. [Online]. Available: https://sonomawinegrape.org/wp-content/uploads/2015 Grape Growing Fact Sheet.pdf. [Accessed: 24-Feb-2019].
[71] P. Reese, “Land values are skyrocketing around Sacramento. See the trend in your community,” The Sacramento Bee, 2019. [Online]. Available: https://www.sacbee.com/news/business/real-estate-news/article225111995.html. [Accessed: 01-Mar-2019].
[72] Acre Value, “Farmland Values, Soil Survey & GIS Maps | AcreValue,” 2019. [Online]. Available: https://www.acrevalue.com/sales/CA/Sonoma/. [Accessed: 01-Mar-2019].
[73] A. C. Goodrich, M. Woodhouse, and T. James, “Solar PV Manufacturing Cost Model Group : Installed Solar PV System Prices,” p. 15, 2011.
[74] CanadianSolar, “KuMax HIGH EFFICIENCY POLY MODULE CS3U-350|355|360P (1000 V / 1500 V),” 2018. [Online]. Available: https://www.canadiansolar.com/downloads/datasheets/en/new/2018-4-16_Ku-v5.571/Canadian_Solar-Datasheet-_KuMax_MBB_5BB__CS3U-P_High_Efficiency_v5.571_EN.pdf. [Accessed: 19-Jan-2019].
[75] Delta, “Delta M80U,” 2017. [Online]. Available:
-82-
www.deltaww.com/filecenter/.../Delta_M80U_DS_20170522.pdf%0A. [Accessed: 19-Jan-2019].
[76] University of Oregon Solar Radiation Monitoring Laboratory, “Sun path chart program,” 2007. [Online]. Available: http://solardat.uoregon.edu/SunChartProgram.html. [Accessed: 19-Jan-2019].
-86-
Appendix D: Floating PV projects
Table 8: Floating PV projects around the world
Year Size (MWp) Location Country Source
2018 0,1 Lismore Australia (1)
2017 1 Geer Belgium (2)
2017 0,304 Cristalina, GO Brazil (3)
2018 5 Balbina, AM Brazil (4)
2019 5 Sobradinho, BA Brazil (5)
2019 0,084 Los Bronces Chile (6)
2015 8 Hebei China (7)
2016 20 Huainan City China (8)
2017 20 Hangzhou China (9)
2017 150 Huainan City China (10)
2017 40 Suixi, Anhui China (11)
2017 40 Anhui China (12)
2018 130 Fuyang China (13)
2018 100 Weishan China (14)
2018 40 Huabei China (15)
2018 0,09936 Medellin Colombia (16)
2019 17 Piolenc France (17)
2014 0,01 Rajarhat India (18)
2017 0,5 Banasura Sagar, Wayanad, Karala India (19)
2018 2 Greater Visakhapatnam India (20)
2011 0,5 Bubano Italy (21)
2013 1,2 Okegawa, Saitama Japan (22)
2015 2 Shiroishi Saga Japan (23)
2015 1 Daiwa Lease, Osaka Japan (24)
2015 1,153 Ranzan-machi, Saitama Japan (25)
2016 2,009 Kato City Yado Oike Japan (26)
2016 7,5 Kawajima-machi, Saitama Japan (27)
2017 2,4 Sanuki, Kagawa Japan (28)
2017 1,52 Sanuki, Kagawa Japan (29)
2017 2,4 Noma Ike Japan (30)
2017 1,5 Mita Kannabe Japan (31)
2017 0,318 Ranzan-machi, Saitama Japan (32)
2018 1,24 Hyogo Japan (33)
2018 13,7 Yamakura Dam Japan (34)
2019 0,906 Kansai, Anazawa Japan (35)
2019 1,57 Kansai, Uozumi Japan (36)
-87-
2019 0,544 Kansai, Kobayashi Japan (37)
2016 0,108 Sepang, Sungai Lubu Malaysia (38)
2017 0,78 Texel Island Netherlands (39)
2018 1,845 Lingewaard Netherlands (40)
2017 0,22 Alto do Rabagão Portugal (41)
2016 2 Boryeong Dam South Korea (42)
2018 3,5 Dangjin South Korea (43)
2018 18,7 Gunsam South Korea (44)
2017 0,392 Murcia Spain (45)
2014 0,2 Berkshire UK (46)
2016 6,3 Walton-on-Thames UK (47)
2016 2,991 Tameside UK (48)
2008 0,206 Kenwood, CA USA (49)
2008 0,03039 Sonoma, CA USA (50)
2012 0,112 Short Hills, NJ USA (51)
2016 0,005 Orlando, FL USA (52)
2016 4,4 Sayreville, NJ USA (53)
2016 0,01 Oakville, CA USA (54)
2017 0,0315 Orlando, FL USA (55)
2018 0,075 Walden, CO USA (56)
2018 0,252 Kelseyville, CA USA (57)
2019 47,5 Binh Thuan Vietnam (58)
-88-
Table 9: Sources for the FPV projects costs
Source Link
(1) https://onestepoffthegrid.com.au/lismore-opens-australias-largest-floating-solar-farm-add-storage/
(2) https://perpetum.be/track-record/#hesbaye
(3) https://economia.estadao.com.br/noticias/mercados,com-painel-solar-conta-de-luz-caiu-de-r-600-para-r-40,70002379090
(4) https://tronicasolar.com.br/2018/02/28/usina-solar-fotovoltaica-flutuante-comeca-a-ser-produzida-na-regiao-do-amazonas/
(5) http://agenciabrasil.ebc.com.br/geral/noticia/2018-11/chesf-inicia-estudo-com-plataformas-solares-no-lago-de-sobradinho
(6) https://www.irishexaminer.com/breakingnews/world/chile-tests-floating-solar-panels-to-power-mine-and-save-water-911241.html
(7) https://renewablesnow.com/news/guangdong-east-builds-8-mwp-floating-pv-system-in-hebei-491707/
(8) https://renewablesnow.com/news/xinyi-solar-plugs-in-20-mw-floating-pv-park-in-china-519795/
(9) https://www.youtube.com/watch?v=R2aQ9fj4xzM
(10) https://www.pv-magazine.com/2017/12/12/chinas-three-gorges-connects-part-of-150-mw-floating-solar-plant/
(11) https://d1jiktx90t87hr.cloudfront.net/449/wp-content/uploads/sites/2/2018/06/Ted-Feierstein-Overview-of-Trina-Solar’s-Experience-in-Floating-Solar.pdf
(12) World Bank Group, ESMAP, and SERIS, “Where sun meets water floating solar market report,” 2018.
(13) https://www.trinasolar.com/en-glb/resources/success-stories/Fuyangshi-yishanggucheng
(14) http://en.sepco1.com/art/2018/5/21/art_3070_202995.html
(15) https://www.trinasolar.com/en-glb/resources/success-stories/Anhui-huaibei-caimeishenxianqu
(16) https://www.larepublica.co/empresas/la-compania-epm-invirtio-cerca-de-800-millones-en-piloto-de-parque-solar-flotante-2715816
(17) https://www.bfmtv.com/economie/les-premieres-centrales-solaires-flottantes-se-jettent-a-l-eau-en-france-1628872.html
(18) https://www.masterbuilder.co.in/countrys-first-floating-solar-plant-come-rajarhat-kolkata/
(19) https://www.ourenergypolicy.org/wp-content/uploads/2018/05/IEEFA-Global-Solar-Report-May-2018.pdf
(20) https://mercomindia.com/india-largest-2mw-floating-solar-vishakhapatnam/
(21) https://www.leggilanotizia.it/2011/05/02/un-impianto-fotovoltaico-galleggiante/
(22) https://renewablesnow.com/news/west-holding-switches-on-15-mw-water-based-pv-plant-in-japan-report-366977/
(23) World Bank Group, ESMAP, and SERIS, “Where sun meets water floating solar market report,” 2018.
(24) https://www.eco-business.com/news/floating-solar-project-completed-in-japan/
(25) https://tech.nikkeibp.co.jp/dm/atclen/news_en/15mk/030101940/?ST=msbe?ST=msbe&P=2
-89-
(26) https://japantoday.com/category/tech/comsys-starts-up-floating-solar-power-plant-in-hyogo
(27) https://tech.nikkeibp.co.jp/dm/atclen/news_en/15mk/013000343/?ST=msbe
(28) https://renewablesnow.com/news/japan-asia-investment-completes-24-mw-floating-pv-system-623902/
(29) https://renewablesnow.com/news/japan-asia-investment-completes-15-mw-floating-pv-system-584197/
(30) World Bank Group, ESMAP, and SERIS, “Where sun meets water floating solar market report,” 2018.
(31) World Bank Group, ESMAP, and SERIS, “Where sun meets water floating solar market report,” 2018.
(32) https://grendz.com/pin/6121/
(33) https://www.renewablesnow.com/news/to-the-point-japans-noritz-completes-124-mw-floating-solar-array-in-hyogo-606166/
(34) World Bank Group, ESMAP, and SERIS, “Where sun meets water floating solar market report,” 2018.
(35) https://renewablesnow.com/news/japans-taiyo-completes-3-mw-of-floating-pv-plants-in-kansai-region-642094/
(36) https://renewablesnow.com/news/japans-taiyo-completes-3-mw-of-floating-pv-plants-in-kansai-region-642094/
(37) https://renewablesnow.com/news/japans-taiyo-completes-3-mw-of-floating-pv-plants-in-kansai-region-642094/
(38) http://www.therakyatpost.com/business/2015/06/04/tnb-explores-floating-solar-power-generation/
(39) Company's private source
(40) https://www.ing.com/Newsroom/All-news/Netherlands-home-to-first-floating-solar-park-in-Europe.htm
(41) https://portugal.edp.com/pt-pt/noticias/2017/12/04/projeto-pioneiro-de-solar-fotovoltaico-flutuante-supera-expetativas
(42) https://www.youtube.com/watch?v=VCooH28KPmY
(43) http://www.chungnamilbo.com/news/articleView.html?idxno=469781
(44) https://news.naver.com/main/read.nhn?mode=LSD&mid=sec&sid1=101&oid=011&aid=0003439955
(45) https://www.energias-renovables.com/fotovoltaica/lorca-estrena-su-gran-flotador-fotovoltaico-20170919
(46) https://inhabitat.com/the-uks-first-floating-solar-farm-unveiled-in-berkshire/
(47) https://www.power-technology.com/features/featuresunshine-on-the-water-floating-solar-arrays-gather-steam-4912782/
(48) https://www.solarpowerportal.co.uk/news/united_utilities_doubles_down_on_floating_pv_with_second_reservoir_install
(49) https://www.nytimes.com/2016/05/24/science/solar-power-floating-on-water.html
(50) Company's private source
(51) E-mail and https://pr.amwater.com/static-files/ed2daf12-4066-4220-a95f-2e9ebf6219cd
-90-
(52) http://digitaledition.orlandosentinel.com/tribune/article_popover.aspx?guid=26648912-0bf9-4bcd-a20c-3d67090dab7a
(53) https://www.rettew.com/project/sayreville-solar-array-design/
(54) Company's private source
(55) https://www.ouc.com/about-ouc/news/2017/02/08/ouc-to-build-rare-floating-solar-array-tomorrow
(56) https://denver.cbslocal.com/2018/10/19/walden-solar-power-floating-panels/
(57) Company's private source
(58) https://renewablesnow.com/news/good-news-for-475-mwp-floating-pv-project-in-vietnam-556764/